| { | |
| "paper_id": "2018", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T07:26:59.688387Z" | |
| }, | |
| "title": "Framework using LSTM Forget Gate For Autism Recognition", | |
| "authors": [ | |
| { | |
| "first": "\u5289\u4e8e\u78a9", | |
| "middle": [ | |
| "\uf02a" | |
| ], | |
| "last": "\u3001\u9673\u91d1\u535a", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "\uf02a", | |
| "middle": [], | |
| "last": "\u3001\u9ad8\u6dd1\u82ac", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "\uf02b", | |
| "middle": [], | |
| "last": "\u3001\u674e\u7948\u5747 \uf02a", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Yu-Shuo", | |
| "middle": [], | |
| "last": "Liu", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Chin-Po", | |
| "middle": [], | |
| "last": "Chen", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Susan", | |
| "middle": [], | |
| "last": "Shur", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Fen", | |
| "middle": [], | |
| "last": "Gau", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Chi-Chun", | |
| "middle": [], | |
| "last": "Lee", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "cclee@ee.nthu.edu.tw" | |
| }, | |
| { | |
| "first": "\uf02a", | |
| "middle": [], | |
| "last": "\u570b\u7acb\u6e05\u83ef\u5927\u5b78\u96fb\u6a5f\u5de5\u7a0b\u5b78\u7cfb", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "Autistic children are less able to tell a fluent story than typical children, so measuring verbal fluency becomes an important indicator when diagnosing autistic children. Fluency assessment, however, needs time-consuming manual tagging, or using expert specially designed characteristics as indicators, therefore, this study proposes a coherence representation learned by directly data-driven architecture, using the forget gate of long short-term memory model to export lexical coherence representation, at the same time, we also use the ADOS coding related to the evaluation of narration to test our proposed representation. Our proposed lexical coherence representation performs high accuracy of 92% on the task of identifying children with autism from typically development. Comparing with the traditional measurement of grammar, word frequency, and latent semantic analysis model, there is a significant improvement. This paper also further randomly shuffles the word order and sentence order, making the typical child's story content become disfluent. By visualizing the data samples after dimension reduction, we further observe the distribution of these fluent, disfluent, and those artificially disfluent data samples. We found the artificially disfluent typical samples would move closer to disfluent autistic samples which prove that our extracted features contain the concept of coherency.", | |
| "pdf_parse": { | |
| "paper_id": "2018", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "Autistic children are less able to tell a fluent story than typical children, so measuring verbal fluency becomes an important indicator when diagnosing autistic children. Fluency assessment, however, needs time-consuming manual tagging, or using expert specially designed characteristics as indicators, therefore, this study proposes a coherence representation learned by directly data-driven architecture, using the forget gate of long short-term memory model to export lexical coherence representation, at the same time, we also use the ADOS coding related to the evaluation of narration to test our proposed representation. Our proposed lexical coherence representation performs high accuracy of 92% on the task of identifying children with autism from typically development. Comparing with the traditional measurement of grammar, word frequency, and latent semantic analysis model, there is a significant improvement. This paper also further randomly shuffles the word order and sentence order, making the typical child's story content become disfluent. By visualizing the data samples after dimension reduction, we further observe the distribution of these fluent, disfluent, and those artificially disfluent data samples. We found the artificially disfluent typical samples would move closer to disfluent autistic samples which prove that our extracted features contain the concept of coherency.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "\u81ea\u9589\u75c7\u662f\u4e00\u7a2e\u666e\u904d\u4e14\u53c8\u75c5\u75c7\u8907\u96dc\u7684\u7cbe\u795e\u75c7\u5019\u7fa4\uff0c\u4e0d\u540c\u65bc\u4e00\u822c\u8f03\u55ae\u7d14\u7684\u7cbe\u795e\u75be\u75c5\uff0c\u6bcf\u500b\u81ea \u9589\u75c7\u60a3\u8005\u4e4b\u9593\u7684\u75c7\u72c0\u7570\u8cea\u6027\u975e\u5e38\u7684\u9ad8\uff0c\u6b64\u4e00\u7279\u6027\u4e5f\u589e\u52a0\u4e86\u8a3a\u65b7\u7684\u56f0\u96e3\u6027\uff0c\u56e0\u6b64\u8a31\u591a\u7814\u7a76 \u81f4\u529b\u65bc\u8a2d\u8a08\u51fa\u5404\u7a2e\u8a55\u91cf\u7684\u6a19\u6e96\u8207\u7279\u5fb5\u4f86\u505a\u70ba\u4eba\u5de5\u81ea\u9589\u75c7\u8a3a\u65b7\u6216\u8005\u81ea\u52d5\u5316\u81ea\u9589\u75c7\u8a3a\u65b7\u7684\u6a19 \u6e96\uff0c\u6b64\u985e\u7814\u7a76\u5c0d\u65bc\u81ea\u9589\u75c7\u7684\u5373\u65e9\u6cbb\u7642\u8207\u7e2e\u77ed\u8a3a\u65b7\u6642\u7a0b\u6709\u6975\u5927\u5e6b\u52a9\uff0c\u4e5f\u56e0\u6b64\u5728\u4eba\u985e\u884c\u70ba\u8a0a \u865f\u8655\u7406\u7684\u9818\u57df\u88e1\u8d8a\u767c\u84ec\u52c3\u767c\u5c55\u3002\u904e\u53bb\u8a31\u591a\u7814\u7a76\u5df2\u91dd\u5c0d\u81ea\u9589\u75c7\u8207\u5178\u578b\u5b69\u7ae5\u7684\u5404\u7a2e\u4e0d\u540c\u6a21\u614b \u884c\u70ba\u9032\u884c\u904e\u5206\u6790\uff0c\u4f8b\u5982\u5229\u7528\u81c9\u90e8\u4e0d\u540c\u7684\u8868\u60c5\u53cd\u61c9\u7279\u5fb5\u4f86\u8fa8\u8b58\u81ea\u9589\u75c7\u7684\u7814\u7a76 (Liu, Li & Yi, 2016) \uff0c\u4ee5\u53ca\u5229\u7528\u5b69\u7ae5\u4e4b\u9593\u4e0d\u540c\u7684\u8a9e\u97f3\u8868\u5fb5 (Marchi et al., 2015; Bone et al., 2014) \uff0c\u8af8\u5982\u8072 \u8abf\u6216\u767c\u97f3\u54ac\u5b57\u7b49\u7b49\uff0c\u4e5f\u6709\u5229\u7528\u5b69\u7ae5\u6558\u4e8b\u7684\u6587\u5b57\u5167\u5bb9\u53bb\u5206\u6790\u8a9e\u6cd5\u548c\u7528\u5b57\u4f5c\u70ba\u5206\u6790\u81ea\u9589\u75c7\u8fa8 \u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u4e4b\u5fd8\u8a18\u9598\u63d0\u53d6\u8a9e\u610f\u6d41\u66a2\u5ea6\u4e4b\u67b6\u69cb\u4ee5\u81ea\u9589\u75c7\u5c0f\u5b69\u8aaa\u6545\u4e8b\u70ba\u4f8b 21 \u8b58\u7279\u5fb5\u7684\u7814\u7a76 (Regneri & King, 2016; Chorianopoulou et al., 2017) \u3002\u6211\u5011\u5728\u6b64\u7bc7\u8ad6\u6587\u4e2d\u63d0\u51fa \u4e00\u7a2e\u8cc7\u6599\u5c0e\u5411\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u63d0\u53d6\u67b6\u69cb\uff0c\u9019\u500b\u67b6\u69cb\u80fd\u5920\u5229\u7528\u81ea\u9589\u75c7\u4ee5\u53ca\u5178\u578b\u5b69\u7ae5\u8aaa\u6545 \u4e8b\u7684\u8cc7\u6599\u8fa8\u8b58\u51fa\u8aaa\u6545\u4e8b\u7684\u8cc7\u6599\u4f86\u81ea\u65bc\u81ea\u9589\u75c7\u5b69\u7ae5\u6216\u5178\u578b\u5b69\u7ae5\u3002 \u904e\u53bb\u5c0d\u65bc\u81ea\u9589\u75c7\u5b69\u7ae5\u8aaa\u6545\u4e8b\u7684\u5206\u6790\u6642\u5e38\u4f7f\u7528\u8a31\u591a\u4e0d\u540c\u7684\u95dc\u9375\u5b57\u7279\u5fb5\u6216\u662f\u4e3b\u984c\u7279\u5fb5\uff0c \u4f86\u89c0\u5bdf\u81ea\u9589\u75c7\u8207\u5178\u578b\u5b69\u7ae5\u7684\u5dee\u7570 (Rouhizadeh, Prud'Hommeaux, Roark & Van Santen, 2013; Rouhizadeh, Sproat, & Van Santen, 2015 )\u3002\u8a31\u591a\u7814\u7a76\u66f4\u6307\u51fa\u76f8\u5c0d\u65bc\u5178\u578b\u5b69\u7ae5\uff0c\u81ea\u9589\u75c7\u5b69\u7ae5 \u8f03\u7121\u6cd5\u8aaa\u51fa\u4e00\u500b\u5b8c\u6574\u7684\u8a9e\u6cd5\u6d41\u66a2\u7684\u6545\u4e8b (Losh & Gordon, 2014) \uff0c\u6216\u662f\u81ea\u7136\u5730\u63cf\u8ff0\u6545\u4e8b\u767c\u5c55 \u5167\u5bb9\u7684\u56e0\u679c\u95dc\u4fc2\u3002\u5206\u6790\u986f\u793a\uff0c\u81ea\u9589\u75c7\u5c0f\u5b69\u63cf\u8ff0\u6545\u4e8b\u6642\uff0c\u50be\u5411\u65bc\u628a\u6545\u4e8b\u4e2d\u7684\u4e00\u500b\u500b\u5143\u7d20\u3001 \u4e8b\u4ef6\u55ae\u7368\u63cf\u8ff0\uff0c\u7f3a\u4e4f\u6d41\u66a2\u7684\u4e32\u63a5\u6216\u6558\u8ff0\u5404\u500b\u60c5\u7bc0\u4e4b\u9593\u7684\u56e0\u679c\u95dc\u4fc2 (Capps, Losh, & Thurber, 2000; Tager-Flusberg, 1995 ) \u3002 \u800c \u8a9e\u6cd5\u8868 \u9054 \u7684\u4e0d \u6d41\u66a2 \u73fe \u8c61\u4e5f \u5df2\u6709 \u7814 \u7a76\u89c0 \u5bdf\u8b49 \u5be6 (Diehl, Bennetto & Young, 2006) (Fadaee, Bisazza & Monz, 2017; Hannun et al., 2014) ", | |
| "cite_spans": [ | |
| { | |
| "start": 228, | |
| "end": 248, | |
| "text": "(Liu, Li & Yi, 2016)", | |
| "ref_id": "BIBREF15" | |
| }, | |
| { | |
| "start": 266, | |
| "end": 287, | |
| "text": "(Marchi et al., 2015;", | |
| "ref_id": "BIBREF18" | |
| }, | |
| { | |
| "start": 288, | |
| "end": 306, | |
| "text": "Bone et al., 2014)", | |
| "ref_id": "BIBREF0" | |
| }, | |
| { | |
| "start": 394, | |
| "end": 416, | |
| "text": "(Regneri & King, 2016;", | |
| "ref_id": "BIBREF22" | |
| }, | |
| { | |
| "start": 417, | |
| "end": 445, | |
| "text": "Chorianopoulou et al., 2017)", | |
| "ref_id": "BIBREF5" | |
| }, | |
| { | |
| "start": 577, | |
| "end": 630, | |
| "text": "(Rouhizadeh, Prud'Hommeaux, Roark & Van Santen, 2013;", | |
| "ref_id": "BIBREF23" | |
| }, | |
| { | |
| "start": 631, | |
| "end": 669, | |
| "text": "Rouhizadeh, Sproat, & Van Santen, 2015", | |
| "ref_id": "BIBREF24" | |
| }, | |
| { | |
| "start": 711, | |
| "end": 732, | |
| "text": "(Losh & Gordon, 2014)", | |
| "ref_id": "BIBREF17" | |
| }, | |
| { | |
| "start": 814, | |
| "end": 844, | |
| "text": "(Capps, Losh, & Thurber, 2000;", | |
| "ref_id": "BIBREF3" | |
| }, | |
| { | |
| "start": 845, | |
| "end": 865, | |
| "text": "Tager-Flusberg, 1995", | |
| "ref_id": "BIBREF26" | |
| }, | |
| { | |
| "start": 902, | |
| "end": 933, | |
| "text": "(Diehl, Bennetto & Young, 2006)", | |
| "ref_id": "BIBREF6" | |
| }, | |
| { | |
| "start": 934, | |
| "end": 964, | |
| "text": "(Fadaee, Bisazza & Monz, 2017;", | |
| "ref_id": "BIBREF7" | |
| }, | |
| { | |
| "start": 965, | |
| "end": 985, | |
| "text": "Hannun et al., 2014)", | |
| "ref_id": "BIBREF10" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "\u7dd2\u8ad6 (Introduction)", | |
| "sec_num": "1." | |
| }, | |
| { | |
| "text": "\u3002\u5728\u9019\u7bc7\u6587\u7ae0\u4e2d,\u6211\u5011\u4f7f\u7528 \u4e00\u500b\u7c21\u55ae\u7684\u6ed1\u52d5\u8996\u7a97\u65b9\u6cd5\u4f86\u57f7\u884c\u8cc7\u6599\u589e\u52a0\u5169\u500b\u8cc7\u6599\u96c6\u3002\u6211\u5011\u7528 n \u500b\u53e5\u5b50\u4f5c\u70ba\u4e00\u7b46\u8cc7\u6599\uff0c \u4e26\u4f7f\u7528\u8207\u6574\u500b\u6587\u6a94\u76f8\u540c\u6a19\u7c64\u7684\u6a23\u672c\uff0c\u6211\u5011\u6bcf\u6b21\u79fb\u52d5\u4e00\u500b\u53e5\u5b50\u4f86\u751f\u6210\u53e6\u4e00\u500b n \u53e5\u7684\u8cc7\u6599\u6a23 \u672c\u3002\u5be6\u9a57\u4e5f\u6bd4\u8f03\u4e86\u7531\u5404\u7a2e\u4e0d\u540c\u500b\u53e5\u5b50\u6578\u7d44\u6210\u7684\u8cc7\u6599\u6a23\u672c\u7684\u7d50\u679c\uff0c\u5206\u5225\u8a66\u9a57\u4e86 n \u70ba 1\uff0c3\uff0c 5\uff0c7\uff0c9 \u7684\u60c5\u6cc1\u3002 2.2.2 \u4e2d\u6587\u6587\u5b57\u65b7\u8a5e (Chinese Word Segmentation) \u8fd1\u5e74\u4f86\u7684\u6587\u5b57\u76f8\u95dc\u7814\u7a76\u4e2d\uff0c\u90fd\u4f7f\u7528\u6587\u5b57\u5411\u91cf\u7a7a\u9593\u4f5c\u70ba\u666e\u904d\u516c\u8a8d\u7684\u8a9e\u610f\u8868\u5fb5\uff0c\u4f46\u4e0d\u540c\u65bc\u82f1 \u6587\uff0c\u4e2d\u6587\u5b57\u5728\u66f8\u5beb\u4e0a\uff0c\u4e26\u6c92\u6709\u5c07\u8a5e\u8a9e\u8a5e\u4e4b\u9593\u9694\u958b\uff0c\u56e0\u6b64\u5728\u9032\u5165\u6587\u5b57\u5411\u91cf\u7684\u8a13\u7df4\u4e4b\u524d\uff0c\u5fc5 \u9808\u5148\u7d93\u904e\u65b7\u8a5e\u7684\u7a0b\u5f0f\uff0c\u9019\u88e1\u6211\u5011\u4f7f\u7528\u7d50\u5df4\u65b7\u8a5e(Sun, 2012)\u9019\u500b\u5de5\u5177\u4f86\u65b7\u8a5e\u3002\u7d50\u5df4\u4f7f\u7528\u6982 \u7387 \u8a9e \u8a00 \u6a21 \u578b \u4f86 \u627e \u51fa \u8a72 \u53e5 \u70ba \u6700 \u53ef \u80fd \u7684 \u8a5e \u7d44 \u5408 \u3002 \u5f9e \u7d71 \u8a08 \u8cc7 \u6599 , \u8f38 \u5165 \u5b57 \u4e32 \u5206 \u5272 \u6a21 \u578b \u7684 , ,\u2026 , ,\u548c\u8f38\u51fa , ,\u2026 , , \u5176\u4e2d \u3002\u5c0d\u65bc\u4e00\u500b\u5b57\u4e32 C\uff0c\u5c07\u6703\u6709\u4e0d\u6b62\u4e00\u7a2e\u53ef \u80fd\u7684\u5206\u5272\u8f38\u51fa S\u3002\u5206\u5272\u6a21\u578b\u7684\u4efb\u52d9\u662f\u627e\u5230\u6982\u7387\u6700\u9ad8\u7684\u8f38\u51fa S \u96c6\u5408\u3002 Seg c argmax \u2208 | argmax \u2208 | \u8209\u4f8b\u4f86\u8aaa: C: \u9752\u86d9\u7d1b\u7d1b\u53c8\u518d\u5ea6\u56de\u5230\u6c34\u88e1 S1: \u9752\u86d9 / \u7d1b\u7d1b / \u53c8\u518d / \u5ea6\u56de / \u5230 / \u6c34\u88e1 S2: \u9752\u86d9 / \u7d1b\u7d1b / \u53c8 / \u518d\u5ea6 / \u56de\u5230 / \u6c34\u88e1 S1 \u548c S2", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "\u7dd2\u8ad6 (Introduction)", | |
| "sec_num": "1." | |
| }, | |
| { | |
| "text": "EQUATION", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [ | |
| { | |
| "start": 0, | |
| "end": 8, | |
| "text": "EQUATION", | |
| "ref_id": "EQREF", | |
| "raw_str": "\u6211\u5011\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u9700\u8981\u5c07\u8aaa\u6545\u4e8b\u6a23\u672c\uff0c\u6295\u5c04\u5230\u9577\u77ed\u671f\u8a18\u61b6\u795e\u7d93\u7db2\u7d61\u5f8c\uff0c\u63d0\u53d6\u907a\u5fd8\u9598\u7684\u8f38 \u51fa \u3002 \u56e0 \u6b64 \u5728 \u672c \u7bc0 \u4e2d \uff0c \u6211 \u5011 \u5c07 \u9996 \u5148 \u7c21 \u8981 \u5730 \u63cf \u8ff0 \u9577 \u77ed \u671f \u8a18 \u61b6 \u795e \u7d93 \u7db2 \u7d61 (Hochreiter & Schmidhuber, 1997)\u3002\u9577\u77ed\u671f\u8a18\u61b6\u795e\u7d93\u7db2\u7d61\u662f\u4e00\u7a2e\u5b9a\u5411\u6642\u9593\u5e8f\u5217\u795e\u7d93\u7db2\u8def\u3002\u9577\u77ed\u671f\u8a18\u61b6\u795e \u7d93\u7db2\u7d61\u7684\u6838\u5fc3\u662f\u5176\u4e2d\u5305\u542b\u7684\u8cc7\u8a0a\u72c0\u614b\u66f4\u65b0\u53c3\u6578 : \u2022 , * * \u4f7f\u7528\u9577\u77ed\u671f\u8a18\u61b6\u795e\u7d93\u7db2\u7d61\u5efa\u6a21\u7684\u597d\u8655\u5728\u65bc\u5b83\u80fd\u5920\u8abf\u7bc0\u9577\u671f\u6216\u77ed\u671f\u8a18\u61b6\u9ad4\u4e0a\u4e0b\u6587\u4e2d\u4fdd\u7559\u7684 \u4fe1\u606f\u91cf\u3002\u8abf\u7bc0\u6a5f\u5236\u662f\u4f7f\u7528\u9598\u7684\u7d50\u69cb\u4f86\u5b8c\u6210\u7684\uff0c\u9598\u7684\u7d50\u69cb\u88ab\u8868\u8ff0\u70ba\u6b0a\u91cd\u77e9\u9663\u548c\u555f\u52d5\u51fd\u6578\u3002 \u6bcf\u500b\u9577\u77ed\u671f\u8a18\u61b6\u795e\u7d93\u7db2\u7d61\u6709\u4e09\u500b\u9598 , ,", | |
| "eq_num": ": \u2022 ," | |
| } | |
| ], | |
| "section": "\u7dd2\u8ad6 (Introduction)", | |
| "sec_num": "1." | |
| }, | |
| { | |
| "text": "\u9996\u5148\uff0c\u89c0\u5bdf\u7279\u5fb5\u9078\u64c7\u7684\u7d50\u679c\uff0c\u5728 TFIDF \u65b9\u6cd5\u65b9\u9762\u986f\u793a\u9023\u63a5\u8a5e\u3001\u526f\u8a5e\u548c\u5e38\u898b\u7684\u77ed\u8a9e\u5728\u6f22\u8a9e \u4e2d\uff0c\u5982\"\u7136\u5f8c\"\u3001\"\u9084\u597d\"\u3001\"\u9019\u6a23\"\u3001\"\u4ec0\u9ebc\"\u90fd\u662f\u81ea\u9589\u75c7\u5b69\u7ae5\u7684\u6558\u8ff0\u91cd\u9ede\u983b\u7387\u660e\u986f \u9ad8\u65bc\u5178\u578b\u5b69\u7ae5\u3002\u95dc\u65bc\u6545\u4e8b\u5167\u5bb9\u7684\u5e7e\u500b\u95dc\u9375\u5b57\uff0c\u4f8b\u5982\"\u5c4b\u9802\"\u3001\"\u8001\u5976\u5976\"\u3001\"\u5730\u677f\"\u3001 \"\u4f4f\"\uff0c\u6070\u6070\u76f8\u53cd\uff0c\u65bc TD (Typically developing)\u53d7\u8a66\u8005\u7684\u6558\u8ff0\u983b\u7387\u4e2d\u986f\u8457\u4f86\u7684\u9ad8\u3002\u6b64\u5916\uff0c \u6d41\u66a2\u5ea6\u77e9\u9663\u7684\u91cd\u8981\u7279\u5fb5\u662f\"\u540d\u8a5e\u91cd\u8907\"\u76f8\u9130\u7684\u53e5\u5b50\"\u548c\"\u540d\u8a5e\u662f\u6574\u9ad4\u7684\u91cd\u8907\u6587\u7ae0\"\u76f8\u9130 \u53e5\u7684\u53e5\u5b50\u76f8\u4f3c\u6027\"\u3002\"\u6574\u7bc7\u6587\u7ae0\u7684\u53e5\u5b50\u76f8\u4f3c\"\u3002\u800c\u9019\u4e9b\u767c\u73fe\uff0c\u78ba\u5be6\u7b26\u5408\u81ea\u9589\u75c7\u60a3\u8005\u4f7f\u7528 \u8f03\u591a\u865b\u8a5e\uff0c\u8f03\u7121\u6cd5\u8868\u9054\u6d41\u66a2\u7684\u53e3\u8a9e\u7684\u60c5\u6cc1\uff0c\u4e5f\u5370\u8b49\u904e\u53bb\u7684\u7814\u7a76\u5f97\u5230\u7684\u7d50\u679c\u3002 ", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "\u5be6\u9a57\u7d50\u679c (Experiment Result)", | |
| "sec_num": "3.2" | |
| } | |
| ], | |
| "back_matter": [], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "The psychologist as an interlocutor in autism spectrum disorder assessment: Insights from a study of spontaneous prosody", | |
| "authors": [ | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Bone", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C.-C", | |
| "middle": [], | |
| "last": "Lee", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [ | |
| "P" | |
| ], | |
| "last": "Black", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [ | |
| "E" | |
| ], | |
| "last": "Williams", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Lee", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "P", | |
| "middle": [], | |
| "last": "Levitt", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Narayanan", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2014, | |
| "venue": "Journal of Speech, Language, and Hearing Research", | |
| "volume": "57", | |
| "issue": "4", | |
| "pages": "1162--1177", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1044/2014_JSLHR-S-13-0062" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Bone, D., Lee, C.-C., Black, M. P., Williams, M. E., Lee, S., Levitt, P. & Narayanan, S. (2014). The psychologist as an interlocutor in autism spectrum disorder assessment: Insights from a study of spontaneous prosody. Journal of Speech, Language, and Hearing Research, 57(4), 1162-1177. doi: 10.1044/2014_JSLHR-S-13-0062", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Signal processing and machine learning for mental health research and clinical applications", | |
| "authors": [ | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Bone", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C.-C", | |
| "middle": [], | |
| "last": "Lee", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "T", | |
| "middle": [], | |
| "last": "Chaspari", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Gibson", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Narayanan", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2017, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Bone, D., Lee, C.-C., Chaspari, T., Gibson, J., & Narayanan, S. (2017). Signal processing and machine learning for mental health research and clinical applications [perspectives].", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "The frog ate the bug and made his mouth sad: Narrative competence in children with autism", | |
| "authors": [ | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Capps", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Losh", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Thurber", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2000, | |
| "venue": "Journal of abnormal child psychology", | |
| "volume": "28", | |
| "issue": "2", | |
| "pages": "193--204", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1023/A:1005126915631" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Capps, L., Losh, M., & Thurber, C. (2000). The frog ate the bug and made his mouth sad: Narrative competence in children with autism. Journal of abnormal child psychology, 28(2), 193-204. doi: 10.1023/A:1005126915631", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Chinese sentence similarity computing and appling", | |
| "authors": [ | |
| { | |
| "first": "S.-Y", | |
| "middle": [], | |
| "last": "Cheng", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "T", | |
| "middle": [], | |
| "last": "Liang", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2005, | |
| "venue": "Proceedings of the 17th Conference on Computational Linguistics and Speech Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "113--124", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Cheng, S.-Y. & Liang, T. (2005). Chinese sentence similarity computing and appling. In Proceedings of the 17th Conference on Computational Linguistics and Speech Processing, 113-124.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Engagement detection for children with autism spectrum disorder", | |
| "authors": [ | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Chorianopoulou", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "E", | |
| "middle": [], | |
| "last": "Tzinis", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "E", | |
| "middle": [], | |
| "last": "Iosif", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Papoulidi", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Papailiou", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Potamianos", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2017, | |
| "venue": "Proceedings of 2017 IEEE International Conference on Acoustics, Speech and Signal Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "5055--5059", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1109/ICASSP.2017.7953119" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Chorianopoulou, A., Tzinis, E., Iosif, E., Papoulidi, A., Papailiou, C., & Potamianos, A. (2017). Engagement detection for children with autism spectrum disorder. In Proceedings of 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017), 5055-5059. doi: 10.1109/ICASSP.2017.7953119", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "Story recall and narrative coherence of highfunctioning children with autism spectrum disorders", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [ | |
| "J" | |
| ], | |
| "last": "Diehl", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Bennetto", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "E", | |
| "middle": [ | |
| "C" | |
| ], | |
| "last": "Young", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2006, | |
| "venue": "Journal of abnormal child psychology", | |
| "volume": "34", | |
| "issue": "1", | |
| "pages": "83--98", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1007/s10802-005-9003-x\u5289\u4e8e\u78a9\u7b49" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Diehl, J. J., Bennetto, L., & Young, E. C. (2006). Story recall and narrative coherence of highfunctioning children with autism spectrum disorders. Journal of abnormal child psychology, 34(1), 83-98. doi: 10.1007/s10802-005-9003-x \u5289\u4e8e\u78a9 \u7b49", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "Data augmentation for low-resource neural machine translation", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Fadaee", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Bisazza", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Monz", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2017, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": { | |
| "arXiv": [ | |
| "arXiv:1705.00440" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Fadaee, M., Bisazza, A., & Monz, C. (2017). Data augmentation for low-resource neural machine translation. Retrieved from arXiv preprint arXiv:1705.00440.", | |
| "links": null | |
| }, | |
| "BIBREF8": { | |
| "ref_id": "b8", | |
| "title": "word2vec explained: deriving mikolov et al.'s negative-sampling wordembedding method", | |
| "authors": [ | |
| { | |
| "first": "Y", | |
| "middle": [], | |
| "last": "Goldberg", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "O", | |
| "middle": [], | |
| "last": "Levy", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2014, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": { | |
| "arXiv": [ | |
| "arXiv:1402.3722" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Goldberg, Y. & Levy, O. (2014). word2vec explained: deriving mikolov et al.'s negative-sampling wordembedding method. Retrieved from arXiv preprint arXiv:1402.3722.", | |
| "links": null | |
| }, | |
| "BIBREF9": { | |
| "ref_id": "b9", | |
| "title": "Coh-metrix: Analysis of text on cohesion and language", | |
| "authors": [ | |
| { | |
| "first": "A", | |
| "middle": [ | |
| "C" | |
| ], | |
| "last": "Graesser", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "D", | |
| "middle": [ | |
| "S" | |
| ], | |
| "last": "Mcnamara", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [ | |
| "M" | |
| ], | |
| "last": "Louwerse", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Z", | |
| "middle": [], | |
| "last": "Cai", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2004, | |
| "venue": "Behavior Research Methods", | |
| "volume": "36", | |
| "issue": "2", | |
| "pages": "193--202", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.3758/BF03195564" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Graesser, A. C., McNamara, D. S., Louwerse, M. M., & Cai, Z. (2004). Coh-metrix: Analysis of text on cohesion and language. Behavior Research Methods, 36(2), 193-202. doi: 10.3758/BF03195564", | |
| "links": null | |
| }, | |
| "BIBREF10": { | |
| "ref_id": "b10", | |
| "title": "Deep speech: Scaling up end-to-end speech recognition", | |
| "authors": [ | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Hannun", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Case", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Casper", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Catanzaro", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "G", | |
| "middle": [], | |
| "last": "Diamos", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "E", | |
| "middle": [], | |
| "last": "Elsen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "A", | |
| "middle": [ | |
| "Y" | |
| ], | |
| "last": "\u2026ng", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2014, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": { | |
| "arXiv": [ | |
| "arXiv:1412.5567" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Hannun, A., Case, C., Casper, J., Catanzaro, B., Diamos, G., Elsen, E., \u2026Ng, A. Y. (2014). Deep speech: Scaling up end-to-end speech recognition. Retrieved from arXiv preprint arXiv:1412.5567.", | |
| "links": null | |
| }, | |
| "BIBREF11": { | |
| "ref_id": "b11", | |
| "title": "Long short-term memory", | |
| "authors": [ | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Hochreiter", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Schmidhuber", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1997, | |
| "venue": "Neural computation", | |
| "volume": "9", | |
| "issue": "8", | |
| "pages": "", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1162/neco.1997.9.8.1735" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Hochreiter, S. & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735-1780. doi: 10.1162/neco.1997.9.8.1735", | |
| "links": null | |
| }, | |
| "BIBREF12": { | |
| "ref_id": "b12", | |
| "title": "Discriminative improvements to distributional sentence similarity", | |
| "authors": [ | |
| { | |
| "first": "Y", | |
| "middle": [], | |
| "last": "Ji", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Eisenstein", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "891--896", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Ji, Y. & Eisenstein, J. (2013). Discriminative improvements to distributional sentence similarity. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, 891-896.", | |
| "links": null | |
| }, | |
| "BIBREF13": { | |
| "ref_id": "b13", | |
| "title": "Understanding blackbox predictions via influence functions", | |
| "authors": [ | |
| { | |
| "first": "P", | |
| "middle": [ | |
| "W" | |
| ], | |
| "last": "Koh", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "P", | |
| "middle": [], | |
| "last": "Liang", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2017, | |
| "venue": "Proceedings of the 34th International Conference on Machine Learning", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "1885--1894", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Koh, P. W. & Liang, P. (2017). Understanding blackbox predictions via influence functions. In Proceedings of the 34th International Conference on Machine Learning (PMLR 70), 1885-1894.", | |
| "links": null | |
| }, | |
| "BIBREF14": { | |
| "ref_id": "b14", | |
| "title": "Understanding neural networks through representation erasure", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Li", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "W", | |
| "middle": [], | |
| "last": "Monroe", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Jurafsky", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2016, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": { | |
| "arXiv": [ | |
| "arXiv:1612.08220" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Li, J., Monroe, W., & Jurafsky, D. (2016). Understanding neural networks through representation erasure. Retrieved from arXiv:1612.08220.", | |
| "links": null | |
| }, | |
| "BIBREF15": { | |
| "ref_id": "b15", | |
| "title": "Identifying children with autism spectrum disorder based on their face processing abnormality: A machine learning framework", | |
| "authors": [ | |
| { | |
| "first": "W", | |
| "middle": [], | |
| "last": "Liu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Li", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Yi", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2016, | |
| "venue": "Autism Research", | |
| "volume": "9", | |
| "issue": "8", | |
| "pages": "888--898", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1002/aur.1615" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Liu, W., Li, M. & Yi, L. (2016). Identifying children with autism spectrum disorder based on their face processing abnormality: A machine learning framework. Autism Research, 9(8), 888-898. doi: 10.1002/aur.1615", | |
| "links": null | |
| }, | |
| "BIBREF16": { | |
| "ref_id": "b16", | |
| "title": "ADOS: Autism diagnostic observation schedule", | |
| "authors": [ | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Lord", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Rutter", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "P", | |
| "middle": [ | |
| "C" | |
| ], | |
| "last": "Dilavore", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Risi", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2008, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Lord, C., Rutter, M., Dilavore, P. C., & Risi, S. (2008). ADOS: Autism diagnostic observation schedule. Boston, MA: Hogrefe.", | |
| "links": null | |
| }, | |
| "BIBREF17": { | |
| "ref_id": "b17", | |
| "title": "Quantifying narrative ability in autism spectrum disorder: A computational linguistic analysis of narrative coherence", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Losh", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "P", | |
| "middle": [ | |
| "C" | |
| ], | |
| "last": "Gordon", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2014, | |
| "venue": "Journal of autism and developmental disorders", | |
| "volume": "44", | |
| "issue": "12", | |
| "pages": "3016--3025", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1007/s10803-014-2158-y" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Losh, M. & Gordon, P. C. (2014). Quantifying narrative ability in autism spectrum disorder: A computational linguistic analysis of narrative coherence. Journal of autism and developmental disorders, 44(12), 3016-3025. doi: 10.1007/s10803-014-2158-y.", | |
| "links": null | |
| }, | |
| "BIBREF18": { | |
| "ref_id": "b18", | |
| "title": "Typicality and emotion in the voice of children with autism spectrum condition: Evidence across three languages", | |
| "authors": [ | |
| { | |
| "first": "E", | |
| "middle": [], | |
| "last": "Marchi", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Schuller", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Baron-Cohen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "O", | |
| "middle": [], | |
| "last": "Golan", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "B\u00a8olte", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "P", | |
| "middle": [], | |
| "last": "Arora", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "H\u00a8ab-Umbach", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2015, | |
| "venue": "Proceedings of Sixteenth Annual Conference of the International Speech Communication Association", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "115--119", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Marchi, E., Schuller, B., Baron-Cohen, S., Golan, O., B\u00a8olte, S., Arora, P., & H\u00a8ab-Umbach, R. (2015). Typicality and emotion in the voice of children with autism spectrum condition: Evidence across three languages. In Proceedings of Sixteenth Annual Conference of the International Speech Communication Association (InterSpeech 2015), 115-119.", | |
| "links": null | |
| }, | |
| "BIBREF19": { | |
| "ref_id": "b19", | |
| "title": "Automated Evaluation of Text and Discourse with Coh-Metrix", | |
| "authors": [ | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Mcnamara", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "A", | |
| "middle": [], | |
| "last": "Graesser", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "P", | |
| "middle": [], | |
| "last": "Mccarthy", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Z", | |
| "middle": [], | |
| "last": "Cai", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2014, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1017/CBO9780511894664" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "McNamara, D., Graesser, A., McCarthy, P., & Cai, Z. (2014). Automated Evaluation of Text and Discourse with Coh-Metrix. Cambridge: Cambridge University Press. doi:10.1017/CBO9780511894664", | |
| "links": null | |
| }, | |
| "BIBREF20": { | |
| "ref_id": "b20", | |
| "title": "Distributed representations of words and phrases and their compositionality", | |
| "authors": [ | |
| { | |
| "first": "T", | |
| "middle": [], | |
| "last": "Mikolov", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "I", | |
| "middle": [], | |
| "last": "Sutskever", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Chen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "G", | |
| "middle": [ | |
| "S" | |
| ], | |
| "last": "Corrado", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Dean", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the \u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u4e4b\u5fd8\u8a18\u9598\u63d0\u53d6\u8a9e\u610f\u6d41\u66a2\u5ea6\u4e4b\u67b6\u69cb\u4ee5\u81ea\u9589\u75c7\u5c0f\u5b69\u8aaa\u6545\u4e8b\u70ba\u4f8b 26th International Conference on neural information processing systems(NIPS'13)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "3111--3119", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Mikolov, T., Sutskever, I., Chen, K., Corrado, G. S., & Dean, J. (2013). Distributed representations of words and phrases and their compositionality. In Proceedings of the \u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u4e4b\u5fd8\u8a18\u9598\u63d0\u53d6\u8a9e\u610f\u6d41\u66a2\u5ea6\u4e4b\u67b6\u69cb\u4ee5\u81ea\u9589\u75c7\u5c0f\u5b69\u8aaa\u6545\u4e8b\u70ba\u4f8b 26th International Conference on neural information processing systems(NIPS'13), 3111-3119.", | |
| "links": null | |
| }, | |
| "BIBREF21": { | |
| "ref_id": "b21", | |
| "title": "Behavioral signal processing: Deriving human behavioral informatics from speech and language", | |
| "authors": [ | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Narayanan", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "P", | |
| "middle": [ | |
| "G" | |
| ], | |
| "last": "Georgiou", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the IEEE", | |
| "volume": "101", | |
| "issue": "", | |
| "pages": "1203--1233", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Narayanan, S. & Georgiou, P. G. (2013). Behavioral signal processing: Deriving human behavioral informatics from speech and language. In Proceedings of the IEEE, 101(5), 1203-1233.", | |
| "links": null | |
| }, | |
| "BIBREF22": { | |
| "ref_id": "b22", | |
| "title": "Automated Discourse Analysis of Narrations by Adolescents with Autistic Spectrum Disorder", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Regneri", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "King", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2016, | |
| "venue": "Proceedings of the 7th Workshop on Cognitive Aspects of Computational Language Learning", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "1--9", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.18653/v1/W16-1901" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Regneri, M., & King, D. (2016). Automated Discourse Analysis of Narrations by Adolescents with Autistic Spectrum Disorder. In Proceedings of the 7th Workshop on Cognitive Aspects of Computational Language Learning, 1-9. doi: 10.18653/v1/W16-1901", | |
| "links": null | |
| }, | |
| "BIBREF23": { | |
| "ref_id": "b23", | |
| "title": "Distributional semantic models for the evaluation of disordered language", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Rouhizadeh", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "E", | |
| "middle": [], | |
| "last": "Prud'hommeaux", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Roark", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Van Santen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics (NAACL-HLT)conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "709--714", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Rouhizadeh, M., Prud'Hommeaux, E., Roark, B., & Van Santen, J. (2013). Distributional semantic models for the evaluation of disordered language. In Proceedings of the Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics (NAACL-HLT)conference, 2013, 709-714.", | |
| "links": null | |
| }, | |
| "BIBREF24": { | |
| "ref_id": "b24", | |
| "title": "Similarity measures for quantifying restrictive and repetitive behavior in conversations of autistic children", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Rouhizadeh", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Sproat", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Van Santen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2015, | |
| "venue": "Proceedings of the 2nd Workshop on Computational Linguistics and Clinical Psychology: From Linguistic Signal to Clinical Reality", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "117--123", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Rouhizadeh, M., Sproat, R., & Van Santen, J. (2015). Similarity measures for quantifying restrictive and repetitive behavior in conversations of autistic children. In Proceedings of the 2nd Workshop on Computational Linguistics and Clinical Psychology: From Linguistic Signal to Clinical Reality, 2015, 117-123.", | |
| "links": null | |
| }, | |
| "BIBREF26": { | |
| "ref_id": "b26", | |
| "title": "Once upon a Ribbit: Stories Narrated by Autistic Children", | |
| "authors": [ | |
| { | |
| "first": "H", | |
| "middle": [], | |
| "last": "Tager-Flusberg", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "British Journal of Developmental Psychology", | |
| "volume": "13", | |
| "issue": "1", | |
| "pages": "45--59", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1111/j.2044-835X.1995.tb00663.x" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Tager-Flusberg, H. (1995). Once upon a Ribbit: Stories Narrated by Autistic Children. British Journal of Developmental Psychology, 13(1), 45-59. doi: 10.1111/j.2044-835X.1995.tb00663.x", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "TABREF0": { | |
| "content": "<table><tr><td>\u5289\u4e8e\u78a9 \u7b49 \u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u4e4b\u5fd8\u8a18\u9598\u63d0\u53d6\u8a9e\u610f\u6d41\u66a2\u5ea6\u4e4b\u67b6\u69cb\u4ee5\u81ea\u9589\u75c7\u5c0f\u5b69\u8aaa\u6545\u4e8b\u70ba\u4f8b 23</td></tr><tr><td>2.1.1 \u8cc7\u6599\u5eab\u4e00:\u81ea\u9589\u75c7\u8207\u5178\u578b\u5b69\u7ae5\u8aaa\u6545\u4e8b\u8cc7\u6599\u5eab (Database I) \u5716 1 \u5448\u73fe\u4e86\u7528\u4f86\u63d0\u53d6\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u7684\u67b6\u69cb\u6d41\u7a0b\uff0c\u6d41\u7a0b\u5305\u542b\u4e86\u8cc7\u6599\u64f4\u589e\u3001\u4e2d\u6587\u6587\u5b57\u5411\u91cf</td></tr><tr><td>\u8cc7\u6599\u5eab\u4e00\u7684\u8490\u96c6\u65b9\u5f0f\u662f\u7531\u5c08\u696d\u57f9\u8a13\u904e\u7684\u7814\u7a76\u4eba\u54e1\u5f15\u5c0e\u5f0f\u7684\u5e36\u9818\u81ea\u9589\u75c7\u5b69\u7ae5\u5b8c\u6210\u7684\uff0c\u4f7f\u7528 \u8a13\u7df4\u6a21\u578b\u3001\u8cc7\u6599\u5eab\u4e8c\u4e4b\u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u9810\u8a13\u7df4\u3001\u8cc7\u6599\u5eab\u4e00\u4e4b\u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u9069\u6027\u8a13\u7df4\u4ee5</td></tr><tr><td>\u7684\u6545\u4e8b\u66f8\u540d\u70ba\u760b\u72c2\u661f\u671f\u4e8c\uff0c\u6545\u4e8b\u66f8\u672c\u8eab\u662f\u4e00\u672c\u7e6a\u672c\uff0c\u66f8\u4e2d\u6bcf\u9801\u5167\u5bb9\uff0c\u90fd\u662f\u4e00\u5f35\u5716\u7247\uff0c\u8207 \u53ca\u6700\u7d42\u7531\u907a\u5fd8\u9598\u63d0\u53d6\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u3002</td></tr><tr><td>\u4e00\u53e5\u77ed\u53e5\u6982\u8981\u9019\u9801\u7684\u5167\u5bb9\uff0c\u800c\u5c0f\u5b69\u5b50\u6703\u6839\u64da\u5716\u7247\u5167\u5bb9\u63cf\u8ff0\u6bcf\u4e00\u9801\u7684\u5287\u60c5\uff0c\u5b8c\u6210\u6574\u7bc7\u6545\u4e8b</td></tr><tr><td>\u7684\u8ad6\u8ff0\u3002 2.2.1 \u8cc7\u6599\u64f4\u589e (Data Augmentation)</td></tr><tr><td>\u800c\u8aaa\u6545\u4e8b\u7684\u9019\u500b\u60c5\u5883\uff0c\u4fbf\u662f\u6a19\u6e96\u7684\u81ea\u9589\u75c7\u8a3a\u65b7\u89c0\u5bdf\u91cf\u8868(Lord, Rutter, Dilavore & Risi, \u7814\u7a76\u8b49\u5be6\uff0c\u900f\u904e\u8cc7\u6599\u64f4\u589e\u7684\u65b9\u5f0f\u80fd\u5920\u4f7f\u7684\u795e\u7d93\u7db2\u7d61\u770b\u904e\u66f4\u5ee3\u6cdb\u7684\u8cc7\u6599\u5206\u4f48\uff0c\u9032\u800c\u6709\u52a9\u65bc</td></tr><tr><td>2008)\u4e2d\u7684\u4e00\u500b\u55ae\u5143\uff0c\u6b64\u70ba\u4e00\u570b\u969b\u516c\u8a8d\u7684\u81e8\u5e8a\u81ea\u9589\u75c7\u8a3a\u65b7\u91cf\u8868\uff0c\u6211\u5011\u662f\u8207\u53f0\u5927\u5152\u7ae5\u91ab\u9662\u5171 \u7a69\u5b9a\u795e\u7d93\u7db2\u7d61\u8a13\u7df4\u7684\u6574\u9ad4\u504f\u5dee\u7387\uff0c\u5728\u795e\u7d93\u7db2\u7d61\u7ffb\u8b6f\u8207\u8a9e\u97f3\u8fa8\u8b58\u7684\u7814\u7a76\u4e2d\u90fd\u5df2\u8b49\u5be6\u9019\u500b\u65b9</td></tr><tr><td>\u540c\u5408\u4f5c\u6536\u96c6\u3002\u9019\u500b\u91cf\u8868\u662f\u8a2d\u8a08\u4f86\u89c0\u6e2c\u5b69\u7ae5\u7684\u793e\u4ea4\u4e92\u52d5\u884c\u70ba\uff0c\u56e0\u6b64\u8a2d\u8a08\u4e86\u8a31\u591a\u884c\u70ba\u55ae\u5143\uff0c \u6cd5\u7684\u5e6b\u52a9\u6027</td></tr><tr><td>\u4f86\u89c0\u6e2c\u4e26\u8a55\u91cf\u53d7\u6e2c\u8005\u7684\u884c\u70ba\u8868\u73fe\uff0c\u55ae\u5143\u5305\u522e\u5efa\u69cb\u5f0f\u4f5c\u696d\u3001\u5047\u626e\u904a\u6232\u3001\u5171\u540c\u4e92\u52d5\u5f0f\u904a\u6232\u3001</td></tr><tr><td>\u793a\u7bc4\u4f5c\u696d\u3001\u5716\u7247\u63cf\u8ff0\u3001\u770b\u5716\u8aaa\u6545\u4e8b\u7b49\u7b49\u3002\u800c\u5728\u6211\u5011\u6240\u4f7f\u7528\u7684\u770b\u5716\u8aaa\u6545\u4e8b\u55ae\u5143\u4e2d\uff0c\u65bd\u6e2c\u8005</td></tr><tr><td>\u3002\u800c\u5728\u4e00\u7bc7 ACL2016 \u7684\u7814\u7a76\u4e2d\u66f4\u5229\u7528\u4e86\u8a31\u591a\u95dc\u9375\u5b57\u6027\u8cea\u6a19\u8a3b\u8868 \u6703\u5148\u5f15\u5c0e\u4e00\u6bb5\u5f15\u6587\uff0c\u800c\u5b69\u7ae5\u5247\u6703\u88ab\u8981\u6c42\u6839\u64da\u5f8c\u534a\u90e8\u7684\u7e6a\u672c\u5167\u5bb9\uff0c\u5b8c\u6210\u5f8c\u534a\u90e8\u7684\u6545\u4e8b\uff0c\u6211</td></tr><tr><td>\u9054\u6574\u500b\u6587\u7ae0\u7684\u5287\u60c5\u8d70\u5411\u8207\u5206\u4f48\uff0c\u85c9\u7531\u9019\u500b\u5c0e\u51fa\u7684\u5287\u60c5\u767c\u5c55\u72c0\u6cc1\u4f86\u7576\u4f5c\u8861\u91cf\u5b69\u7ae5\u8aaa\u6545\u4e8b\u7684 \u5011\u7528\u76f8\u540c\u7684\u6d41\u7a0b\u6536\u96c6\u4e86\u5178\u578b\u5b69\u7ae5\u90e8\u5206\u7684\u8aaa\u6545\u4e8b\u8cc7\u6599\uff0c\u9010\u5b57\u7a3f\u7684\u5167\u5bb9\u8209\u4f8b\u5982:</td></tr><tr><td>\u6d41\u66a2\u6027\u3002\u7136\u800c\uff0c\u4ee5\u4e0a\u65b9\u6cd5\u7686\u9808\u4ef0\u8cf4\u7d93\u904e\u8a13\u7df4\u7684\u5c08\u696d\u4eba\u58eb\u6a19\u8a3b\uff0c\u6a19\u8a3b\u5341\u5206\u8cbb\u6642\uff0c\u4e14\u6703\u53d7\u9650 \"\u661f\u671f\u4e8c\u7684\u665a\u4e0a\uff0c\u6709\u4e00\u7fa4\u9752\u86d9\u5728\u6c60\u5858\u4e0a\uff0c\u4ed6\u5011\u7a81\u7136\u4e58\u8457\u8377\u8449\u98db\u4e86\u8d77\u4f86\uff0c\u4ed6\u5011\u98db\u5230\u4e86\u9130 \u65bc\u53ea\u80fd\u5728\u88ab\u8a2d\u8a08\u7684\u60c5\u6cc1\u4e0b\u4f7f\u7528\u3002 \u8fd1\u7684\u5c0f\u93ae\uff0c\u63a5\u4e0b\u4f86\u63db\u4f60\u8aaa\u6545\u4e8b\u2026\u2026\"</td></tr><tr><td>\u56e0\u6b64\uff0c\u6211\u5011\u7684\u7814\u7a76\u4f7f\u7528\u8cc7\u6599\u5c0e\u5411\u7684\u65b9\u6cd5\uff0c\u4ee5\u6587\u5b57\u5411\u91cf\u7576\u4f5c\u8f38\u5165\uff0c\u8f38\u9032\u9577\u77ed\u671f\u8a18\u61b6\u6a21 \u5728\u9019\u500b\u8cc7\u6599\u5eab\u4e2d\u6211\u5011\u7e3d\u5171\u6536\u96c6\u4e86 67 \u4f4d\u5b69\u7ae5\u7684\u8cc7\u6599\u7e3d\u5171\u7d04\u6709 28446 \u5b57\u7684\u8cc7\u6599\u5eab\uff0c\u6bcf \u578b\uff0c\u4f7f\u7528\u671f\u8a13\u7df4\u597d\u7684\u907a\u5fd8\u9598\u4f5c\u70ba\u53c3\u6578\uff0c\u8b93\u6a5f\u5668\u81ea\u5df1\u5b78\u51fa\u8868\u9054\u8cc7\u6599\u4e2d\u542b\u6709\u6d41\u66a2\u5ea6\u7684\u6642\u5e8f\u6027 \u7bc7\u6545\u4e8b\u5e73\u5747\u7d04\u6709 424 \u5b57\uff0c67 \u4f4d\u5b69\u7ae5\u4e2d\u5305\u62ec 31 \u4f4d\u81ea\u9589\u75c7\u5b69\u7ae5(ASD)\u4ee5\u53ca 36 \u4f4d\u5178\u578b\u5b69\u7ae5(TD) \u3002 \u8a9e\u6cd5\u7279\u5fb5\u3002\u5be6\u9a57\u4e2d\uff0c\u6211\u5011\u6bd4\u8f03\u4ee5\u5404\u4e0d\u540c\u7684\u7279\u5fb5\u4f86\u8a13\u7df4\u91dd\u5c0d\u8aaa\u6545\u4e8b\u8cc7\u6599\u5eab\u7684\u81ea\u9589\u75c7\u8fa8\u8b58\u6a21 \u578b\u7684\u6e96\u78ba\u7387\uff0c\u800c\u6211\u5011\u6240\u63d0\u51fa\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u6240\u8a13\u7df4\u7684\u6a21\u578b\u5728\u5be6\u9a57\u4e2d\u9054\u5230 0.92 \u7684\u6e96\u78ba\u7387\uff0c \u8868 1 \u70ba\u8cc7\u6599\u5eab\u76f8\u95dc\u6558\u8ff0\u3002</td></tr><tr><td>\u9ad8\u904e\u65bc\u7531\u77e5\u540d\u7684\u6d41\u66a2\u5ea6\u7279\u5fb5\uff0c\u6d41\u66a2\u77e9\u9663(Graesser, McNamara, Louwerse & Cai, 2004;</td></tr><tr><td>McNamara, Graesser, McCarthy & Cai, 2014)\u6240\u8a13\u7df4\u51fa\u4f86\u7684\u6a21\u578b\u3002\u518d\u8005\uff0c\u555f\u767c\u65bc\u6700\u65b0\u95dc\u65bc</td></tr><tr><td>\u7406\u89e3\u795e\u7d93\u7db2\u8def\u9ed1\u76d2\u5b50\u7684\u7814\u7a76(Li, Monroe & Jurafsky, 2016; Koh & Liang, 2017)\uff0c\u6211\u5011\u89c0\u5bdf</td></tr><tr><td>\u6539\u8b8a\u539f\u59cb\u8cc7\u6599\u7684\u6027\u8cea\u5f8c\u7684\u8a9e\u610f\u6d41\u66a2\u7279\u5fb5\u5206\u4f48\u7684\u6539\u8b8a\uff0c\u4f86\u8a66\u8457\u7406\u89e3\u63a8\u6572\u795e\u7d93\u7db2\u8def\u9019\u500b\u9ed1\u76d2</td></tr><tr><td>\u5b50\u904b\u4f5c\uff0c\u4e26\u9a57\u8b49\u6211\u5011\u5c0e\u51fa\u7684\u8868\u5fb5\u7684\u610f\u7fa9\u3002\u5be6\u9a57\u4e2d\uff0c\u6211\u5011\u5c07\u5178\u578b\u5b69\u7ae5\u6240\u6558\u8ff0\u6545\u4e8b\u7684\u5b57\u5e8f\u548c</td></tr><tr><td>\u53e5\u5e8f\u5206\u5225\u6253\u4e82\uff0c\u4f86\u6a21\u64ec\u4e0d\u540c\u7a0b\u5ea6\u7684\u4e0d\u6d41\u66a2\uff0c\u7d50\u679c\u9a5a\u8a1d\u7684\u767c\u73fe\uff0c\u5728\u4ee5\u6211\u5011\u5c0e\u51fa\u7684\u6d41\u66a2\u5ea6\u8868</td></tr><tr><td>\u5fb5\u7a7a\u9593\u88e1\uff0c\u88ab\u6253\u4e82\u5f8c\u7684\u4e0d\u6d41\u66a2\u5178\u578b\u5c0f\u5b69\u8aaa\u6545\u4e8b\u8cc7\u6599\u5206\u4f48\uff0c\u6703\u5f80\u7121\u6cd5\u8aaa\u51fa\u6d41\u66a2\u6545\u4e8b\u7684\u81ea\u9589</td></tr><tr><td>\u75c7\u5c0f\u5b69\u8aaa\u51fa\u7684\u6545\u4e8b\u8cc7\u6599\u5206\u4f48\u9760\u8fd1\uff0c\u6b64\u7d50\u679c\u9a57\u8b49\u4e86\u6211\u5011\u5c0e\u51fa\u7684\u8868\u5fb5\u7684\u78ba\u542b\u6709\u6d41\u66a2\u5ea6\u7684\u8cc7\u8a0a\u3002</td></tr><tr><td>\u6b64\u7bc7\u8ad6\u6587\u5167\u5bb9\u5b89\u6392\u5982\u4e0b: \u7b2c\u4e8c\u7bc0\u70ba\u8cc7\u6599\u5eab\u4ee5\u53ca\u67b6\u69cb\u4ecb\u7d39\uff0c\u7b2c\u4e09\u7bc0\u70ba\u5be6\u9a57\u7d50\u679c\u5206\u6790\u8207\u5448\u73fe\uff0c</td></tr><tr><td>\u6700\u5f8c\u7b2c\u56db\u7bc0\u70ba\u7d50\u8ad6\u3002</td></tr><tr><td>2. \u7814\u7a76\u65b9\u6cd5 (Research Methodology)</td></tr><tr><td>2.1 \u8cc7\u6599\u5eab (Database)</td></tr><tr><td>\u9019\u7bc7\u8ad6\u6587\u4e2d\uff0c\u6211\u5011\u4f7f\u7528\u5169\u500b\u8cc7\u6599\u5eab\uff0c\u7b2c\u4e00\u500b\u662f\u4e3b\u8981\u7528\u4f86\u9a57\u8b49\u6211\u5011\u6240\u63d0\u53d6\u51fa\u7684\u6d41\u66a2\u5ea6\u7279\u5fb5</td></tr><tr><td>\u8fa8\u8b58\u529b\u7684\u81ea\u9589\u75c7\u8207\u5178\u578b\u5b69\u7ae5\u8aaa\u6545\u4e8b\u8cc7\u6599\u5eab\uff0c\u7b2c\u4e8c\u500b\u5247\u662f\u5f9e\u6536\u9304\u4e86\u8a31\u591a\u5152\u7ae5\u8b80\u7269\u7684\u7ae5\u8a71\u6545</td></tr><tr><td>\u4e8b\u7db2\u7ad9\uff0c\u6240\u722c\u4e0b\u4f86\u7684\u56db\u7a2e\u7ae5\u8a71\u6545\u4e8b\u8cc7\u6599\u5eab\uff0c\u7b2c\u4e8c\u500b\u8cc7\u6599\u5eab\u662f\u7528\u4f86\u9810\u8a13\u7df4\u4f7f\u7528\u3002</td></tr></table>", | |
| "text": "", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF1": { | |
| "content": "<table><tr><td>\u5289\u4e8e\u78a9 \u7b49</td></tr><tr><td>\u6295\u5c04\u5230\u96b1\u85cf\u7684\u7dad\u5411\u91cf\u7a7a\u9593\u6a21\u578b\u4e2d\u7684\u5ea7\u6a19\u4e0a\u3002\u6211\u5011\u53ef\u4ee5\u901a\u904e\u8a08\u7b97\u4e0d\u540c\u8a5e\u9593\u7684\u5354\u8abf\u8ddd\u96e2\u4f86\u767c</td></tr><tr><td>\u73fe\u6bcf\u500b\u8a5e\u4e4b\u9593\u7684\u95dc\u4fc2\u3002Mikolov \u63d0\u51fa\u4e86\u5169\u7a2e\u4e0d\u540c\u7684\u65b9\u6cd5\uff0c\u4e00\u662f\u9023\u7e8c\u7684\u8a5e\u888b(CBOW)\uff0c\u53e6\u4e00</td></tr><tr><td>\u7a2e\u662f Skip-gram (Skip-gram)\u3002\u4ed6\u9084\u63d0\u51fa\u4e86\u5169\u7a2e\u8a08\u7b97\u6548\u7387\u9ad8\u7684\u8fd1\u4f3c\u5c64\u6b21\u5316\u548c\u8ca0\u63a1\u6a23\u3002\u5229\u7528</td></tr><tr><td>\u9019\u5169\u7a2e\u6709\u6548\u7684\u903c\u8fd1\u65b9\u6cd5\uff0c\u8a72\u6a21\u578b\u53ef\u4ee5\u4ee5\u4e00\u7a2e\u66f4\u6709\u6548\u7684\u65b9\u5f0f\u5b78\u7fd2\u6b63\u78ba\u7684\u8868\u793a\u65b9\u6cd5(Goldberg</td></tr><tr><td>& Levy, 2014)\u3002</td></tr><tr><td>\u9019\u88e1\u6211\u5011\u4f7f\u7528 CBOW \u4f86\u69cb\u5efa\u6211\u5011\u7684\u8a5e\u5d4c\u5165\u7db2\u8def\u3002\u4e3b\u8981\u601d\u60f3\u662f\u5229\u7528 2c \u76f8\u9130\u8a5e\u7684\u4e0a\u4e0b</td></tr><tr><td>\u6587\u5411\u91cf\u4f86\u9810\u6e2c\u76ee\u6a19\u8a5e\uff0c\u6a21\u578b\u7d50\u69cb\u5305\u62ec K \u500b\u8a5e\u5f59\u91cf\u5b57\u5178\u3001\u8f38\u5165\u5c64\u3001\u6295\u5f71\u5c64\u548c\u8f38\u51fa\u5c64\u3002\u8f38\u5165</td></tr><tr><td>\u5c64\u7684\u8f38\u5165\u662f\u6bcf\u500b\u76f8\u9130\u5b57\u7684\u7368\u71b1\u7de8\u78bc(one-hot)\u5411\u91cf\u3002\u6211\u5011\u4f7f\u7528 K x V \u8f38\u5165\u5c64\u7dad\u5ea6\u5f97\u5230\u6bcf\u500b</td></tr><tr><td>\u76f8\u9130\u8a5e\u4e0a\u4e0b\u6587\u5411\u91cf w\uff0cw \u7684\u7dad\u5ea6\u70ba\u81ea\u8a02\u5927\u5c0f\uff0c\u800c\u5f8c\u6211\u5011\u5f97\u5230\u9019\u4e9b\u4e0a\u4e0b\u6587\u7684\u5e73\u5747\u5411\u91cf V(w)</td></tr><tr><td>\u50b3\u64ad\u5230\u8f38\u51fa\u5c64\u9810\u6e2c\u76ee\u6a19\u8a5e w\u3002\u6700\u7d42\uff0c\u8a72\u6a21\u578b\u53ef\u4ee5\u4f7f\u7528\u53cd\u5411\u50b3\u64ad\u4f86\u66f4\u65b0\u53c3\u6578\uff0c\u672c\u7bc7\u5be6\u9a57\u4f7f</td></tr><tr><td>\u7528\u8cc7\u6599\u5eab\u4e00\u4ee5\u53ca\u8cc7\u6599\u5eab\u4e8c\u7684\u5168\u90e8\u6587\u6a94\u8a13\u7df4\u51fa\u6587\u5b57\u8f49\u5411\u91cf\u7684\u7db2\u7d61\u6a21\u578b\uff0c\u800c\u8a5e\u983b\u6578\u904e\u4f4e\u7684\u9805</td></tr><tr><td>\u76ee\u6703\u88ab\u8f49\u6210 OOV(out of vocabulary)\u4e4b\u5411\u91cf\u3002</td></tr><tr><td>2.2</td></tr><tr><td>\u4fbf\u662f\u4e0d\u540c\u7684\u53ef\u80fd\u8a5e\u7d44\u5408\uff0c\u6839\u64da\u4e0d\u540c\u7684\u5206\u5272 S\uff0c\u8a5e\u6578 m \u7684\u500b\u6578\u4e0d\u540c\u3002m \u8d8a\u5927\uff0c\u5be6</td></tr><tr><td>\u969b\u89c0\u6e2c\u7684\u53ef\u80fd\u8a5e\u7d44\u6210\u6982\u7387\u8d8a\u5c0f\u3002</td></tr><tr><td>2.2.3 \u4e2d\u6587\u6587\u5b57\u8f49\u5411\u91cf (Chinese Word2Vec)</td></tr><tr><td>\u8207\u8a9e\u97f3\u4fe1\u865f\u548c\u8996\u8a0a\u8a0a\u865f\u4e0d\u540c\uff0c\u6587\u672c\u662f\u96e2\u6563\u7684\uff0c\u4e0d\u662f\u9023\u7e8c\u53ef\u5fae\u7684\u3002\u70ba\u4e86\u4f7f\u5b83\u80fd\u5920\u8f38\u5165\u795e\u7d93</td></tr><tr><td>\u7db2\u8def\uff0c\u6210\u70ba\u66f4\u597d\u7684\u9032\u4e00\u6b65\u5206\u6790\u7684\u8868\u793a\uff0c\u6211\u5011\u4f7f\u7528\u6587\u5b57\u5411\u91cf\u65b9\u6cd5\u5c07\u6211\u5011\u8a9e\u6599\u5eab\u4e2d\u7684\u6bcf\u500b\u8a5e</td></tr><tr><td>\u6295\u5c04\u5230\u4f4e\u7dad\u7a7a\u9593\u7684\u7279\u5b9a\u5354\u8abf\u4e0a\u3002\u56e0\u6b64\uff0c\u6211\u5011\u73fe\u5728\u628a\u96e2\u6563\u6587\u672c\u8f49\u63db\u6210\u9023\u7e8c\u53ef\u5fae\u7684\u8a5e\u5411\u91cf\u3002</td></tr><tr><td>\u6587\u5b57\u5411\u91cf\u7a7a\u9593(Mikolov, Sutskever, Chen, Corrado & Dean,2013)\u662f\u4e00\u7a2e\u8cc7\u6599\u9a45\u52d5\u7684\u5b78</td></tr><tr><td>\u7fd2\u8868\u793a\u5f62\u5f0f\uff0c\u662f\u5f9e\u4e00\u500b\u9f90\u5927\u7684\u901a\u7528\u8a9e\u6599\u5eab\u4e2d\u5b78\u7fd2\u7684\u3002\u5728\u5b78\u7fd2\u904e\u7a0b\u4e4b\u5f8c\uff0c\u6a21\u578b\u5c07\u6bcf\u500b\u55ae\u8a5e</td></tr></table>", | |
| "text": "", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF2": { | |
| "content": "<table><tr><td/><td colspan=\"3\">\u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u4e4b\u5fd8\u8a18\u9598\u63d0\u53d6\u8a9e\u610f\u6d41\u66a2\u5ea6\u4e4b\u67b6\u69cb\u4ee5\u81ea\u9589\u75c7\u5c0f\u5b69\u8aaa\u6545\u4e8b\u70ba\u4f8b \u5289\u4e8e\u78a9 \u7b49 25 \u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u4e4b\u5fd8\u8a18\u9598\u63d0\u53d6\u8a9e\u610f\u6d41\u66a2\u5ea6\u4e4b\u67b6\u69cb\u4ee5\u81ea\u9589\u75c7\u5c0f\u5b69\u8aaa\u6545\u4e8b\u70ba\u4f8b 27</td></tr><tr><td colspan=\"4\">\u6599\u96c6\u4e00\u9032\u884c\u5fae\u8abf\uff0c\u5f97\u5230\u6700\u5f8c\u7684\u7d50\u679c\u3002\u6700\u5f8c\u63d0\u53d6\u6240\u5b78\u7fd2\u7684\u907a\u5fd8\u9598\u77e9\u9663W \uff0c\u63d0\u53d6\u5f8c\uff0c\u6211\u5011\u53ef \u4ee5\u8a08\u7b97\u51fa\uff0c\u5728\u555f\u52d5\u51fd\u6578\u4e4b\u524d\uff0c\u6c92\u6709\u504f\u501a\u9805\u7684\u907a\u5fd8\u9598\u7684\u8f38\u51fa\u503c : \u2022 , \u6d41\u66a2\u5ea6\u77e9\u9663\u88ab\u8a8d\u70ba\u662f\u5c0d\u6587\u672c\u548c\u8a71\u8a9e\u6700\u8a73\u76e1\u7684\u81ea\u52d5\u8a55\u4f30\u4e4b\u4e00\u3002\u5b83\u8a08\u7b97\u66f8\u5beb\u7684\u4e00\u81f4\u6027\uff0c\u4e26\u80fd \u7528\u5728\u5ea6\u91cf\u53e3\u8a9e\u548c\u6558\u5beb\u7684\u5167\u5bb9(McNamara et al., 2014)\u3002\u6e2c\u91cf\u65b9\u6cd5\u5305\u62ec\u76f8\u9130\u53e5\u9593\u7684\u76f8\u4f3c\u5ea6\u8a08 \u7b97\u3001\u76f8\u9130\u53e5\u6216\u6574\u7bc7\u6587\u7ae0\u4e2d\u540d\u8a5e\u91cd\u758a(\u91cd\u8907)\u3001\u6587\u7ae0\u53e5\u5b50\u7d50\u69cb\u7684\u76f8\u4f3c\u5ea6\u7b49\u3002\u4f8b\u5982\uff0c\u76f8\u9130\u53e5\u9593 \u7684\u76f8\u4f3c\u5ea6\u901a\u904e\u6bcf\u53e5\u7684 LSA(\u6f5b\u8a9e\u7fa9\u5206\u6790)\u5411\u91cf\u8a08\u7b97\u3002\u7136\u5f8c\u8a08\u7b97\u5169\u500b LSA \u5411\u91cf\u7684\u9918\u5f26\u76f8\u4f3c 3. \u5be6\u9a57\u8a2d\u8a08\u53ca\u7d50\u679c \u5728\u9019\u7bc7\u7814\u7a76\u4e2d\uff0c\u6211\u5011\u4f7f\u7528\u6211\u5011\u63d0\u51fa\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u8868\u793a\u5728\u5178\u578b\u5b69\u7ae5\u8207\u81ea\u9589\u75c7\u5b69\u7ae5\u4e4b\u9593 \u5c0d\u65bc\u6bcf\u500b\u8cc7\u6599\u6a23\u672c\uff0c\u5373 n \u500b\u53e5\u5b50,\u6bcf\u500b\u53e5\u5b50\u5206\u5225 \u500b\u5b57,\u6703\u6709\u7e3d n x k \u5927\u5c0f\u7684 \u5e8f\u5217\u3002\u6211\u5011 \u5ea6\u3002 \u9032\u884c\u5206\u985e\u3002\u6211\u5011\u6bd4\u8f03\u4e0b\u5217\u65b9\u6cd5: \u5c07\u5b83\u5011\u7de8\u78bc\u70ba\u6bcf\u500b\u8cc7\u6599\u6a23\u672c\u90fd\u56fa\u5b9a\u7dad\u5ea6\u7684\u8868\u793a\u5f62\u5f0f\u3002 \u5716 2 \u70ba\u6d41\u66a2\u5ea6\u77e9\u9663\u7684\u4f7f\u7528\u4ecb\u9762\uff0c\u5c07\u6b32\u8a08\u7b97\u6d41\u66a2\u5ea6\u4e4b\u6587\u7ae0\u6587\u5b57\u5167\u5bb9\u8f38\u5165\u5f8c\uff0c\u4fbf\u53ef\u52fe\u9078 (1) \u8a5e\u983b\u9006\u6587\u6a94\u983b\u7387:\u5171 2034 \u7dad\u7684\u7279\u5fb5\u503c F g , \u2026 , \u5404\u9805\u7279\u5fb5\uff0c\u5f97\u5230\u5404\u7279\u5fb5\u503c\uff0c\u4f8b\u5982\u5716\u4e2d\u52fe\u9078\u7684\u5de6\u908a\u4fbf\u662f\u900f\u904e LSA \u5411\u91cf\u8a08\u7b97\u8a9e\u8a5e\u9593\u8a9e\u7fa9\u7684\u76f8 (2) \u6d41\u66a2\u5ea6\u77e9\u9663:\u5171 20 \u7dad\u7684\u7279\u5fb5\u503c</td></tr><tr><td colspan=\"4\">\u5176\u4e2d g \u8868\u793a 17 \u500b\u7d71\u8a08\u6cdb\u51fd\u6578\uff0c\u6700\u5927\u503c(MAX)\uff0c\u6700\u5c0f\u503c(Min)\uff0c\u5e73\u5747\u6578(Mean)\uff0c\u4e2d\u4f4d\u6578 \u4f3c\u5ea6\uff0c\u800c\u53f3\u908a\u5247\u662f\u55ae\u7d14\u8a08\u7b97\u4e0d\u540c\u8a5e\u6027\u5b57\u8a5e\u91cd\u8907\u7684\u7a0b\u5ea6\u3002 (3) \u9577\u77ed\u671f\u8a18\u61b6\u795e\u7d93\u7db2\u7d61:\u53d6\u51fa\u96b1\u85cf\u5c64\u7684\u8f38\u51fa\u4f86\u7576\u4f5c\u7279\u5fb5\u9032\u884c\u5206\u985e (Medium)\uff0c\u6a19\u6e96\u5dee(Std)\uff0c1 \u767e\u5206\u4f4d\u6578(1%)\uff0c99 \u767e\u5206\u4f4d\u6578(99%)\uff0c99 \u767e\u5206\u4f4d\u6578-1 \u767e\u5206\u4f4d (99%-1%)\u6578\uff0c\u504f\u5ea6(skewness)\uff0c\u5cf0\u5ea6(kurtosis)\uff0c\u6700\u5c0f\u4f4d(min pos)\uff0c\u6700\u5927\u4f4d(max pos)\uff0c\u4e0b\u56db \u5206\u4f4d\u6578(lower quartile)\uff0c\u4e0a\u56db\u5206\u4f4d\u6578(upper quartile)\uff0c\u4e0a\u56db\u5206\u4f4d\u6578\u7bc4\u570d(interquartile range)\uff0c \u51aa(power)\uff0c1 \u5dee\u5206(point difference)\u3002\u9019\u500b F \u4fbf\u662f\u6bcf\u500b\u8cc7\u6599\u6a23\u672c\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u8868\u5fb5\u3002 2.3 \u66a2\u5ea6\u77e9\u9663 (Cohesion Matrix) \u6211\u5011\u4f7f\u7528\u6240\u6709\u7684 20 \u7dad\u7279\u5fb5\uff0c\u901a\u904e\u76f8\u95dc\u6027\u7279\u5fb5\u9078\u64c7\uff0c\u6211\u5011\u5728\u9078\u64c7 25%\u7684\u7279\u6027\u96c6\u6642\u7372 (4) \u8a9e\u610f\u6d41\u66a2\u5ea6\u8868\u5fb5:\u70ba\u672c\u7bc7\u6240\u63d0\u51fa\u4e4b\u8868\u5fb5 \u5f97\u6700\u4f73\u6027\u80fd\u3002\u6240\u9078\u64c7\u7684\u4e94\u500b\u7279\u5fb5\u662f\u5c40\u90e8\u540d\u8a5e\u91cd\u758a\u3001\u5168\u529f\u80fd\u8b8a\u6578\u540d\u7a31\u8a5e\u91cd\u758a\u3001\u5c40\u90e8\u53e5\u5b50\u7d50 \u5be6\u9a57\u4e2d\u4f7f\u7528\u4e86\u5404\u9805\u7279\u5fb5\u4f5c\u70ba\u6d41\u66a2\u5ea6\u7279\u5fb5\u7684\u6bd4\u8f03\uff0c\u5982 Coh-Metrix \u4e3b\u8981\u662f\u5229\u7528\u53e5\u5b50\u4e4b\u9593 \u69cb\u76f8\u4f3c\u3001\u5168\u57df\u53e5\u5b50\u7d50\u69cb\u76f8\u4f3c\u4ee5\u53ca\u53e5\u5b50\u4e2d\u6700\u5c0f\u7684\u55ae\u8a5e\u91cd\u8907\u983b\u7387\u3002 \u7684\u76f8\u4f3c\u5ea6\uff0c\u4f86\u8861\u91cf\u4e00\u7bc7\u6587\u7ae0\u7684\u6d41\u66a2\uff0c\u6240\u4f7f\u7528\u7684\u6982\u5ff5\u662f\uff0c\u7576\u4e00\u7bc7\u6587\u7ae0\u88e1\u76f8\u9130\u7684\u53e5\u5b50\u8a9e\u7fa9\u7a7a \u5c40\u90e8\u540d\u8a5e\u91cd\u758a\u662f\u6307\u8a08\u7b97\u76f8\u9130\u53e5\u5b50\u4e2d\u540c\u4e00\u540d\u8a5e\u7684\u91cd\u8907\u983b\u7387\uff0c\u5168\u529f\u80fd\u8b8a\u6578\u540d\u7a31\u8a5e\u91cd\u758a\u662f \u9593\u76f8\u4f3c\u5ea6\u9ad8\u6642\uff0c\u53ef\u4ee5\u4ee3\u8868\u6558\u8ff0\u9023\u8cab\uff0c\u53cd\u4e4b\u5247\u662f\u4e0d\u6d41\u66a2\u3002\u800c\u900f\u904e TF-IDF \u7684\u65b9\u6cd5\u6709\u6a5f\u6703\u8b93 \u6307\u8a08\u7b97\u540c\u4e00\u540d\u8a5e\u5728\u4e00\u7bc7\u6587\u7ae0\u4e2d\u5343\u8a5e\u7684\u91cd\u8907\u983b\u7387\u3002\u5c40\u90e8\u548c\u5168\u57df\u53e5\u5b50\u7d50\u69cb\u7684\u76f8\u4f3c\u6027\u662f\u6307\u8a08\u7b97 \u6a21\u578b\u900f\u904e\u95dc\u9375\u8a5e\u6578\u53bb\u5b78\u7fd2\uff0c\u662f\u5426\u4e00\u7bc7\u6587\u7ae0\u4e00\u76f4\u4f7f\u7528\u904e\u591a\u91cd\u8907\u7684\u8a5e\u8a9e\uff0c\u9020\u6210\u4e0d\u6d41\u66a2\u7684\u73fe\u8c61\uff0c \u76f8\u9130\u53e5\u5b50\u7d50\u69cb\u7684\u76f8\u4f3c\u6027\uff0c\u8a08\u7b97\u6574\u500b\u53e5\u5b50\u7d50\u69cb\u7684\u76f8\u4f3c\u6027\u3002\u53e5\u5b50\u7d50\u69cb\u7684\u76f8\u4f3c\u6027\u4e00\u822c\u63a1\u7528\u8a08\u7b97 \u800c LSTM \u5247\u662f\u5728\u4f9d\u5e8f\u95b1\u8b80\u904e\u6bb5\u843d\u6587\u5b57\u4e4b\u5f8c\uff0c\u7d66\u51fa\u4e00\u500b\u65b7\u5b9a\u6d41\u66a2\u985e\u6216\u4e0d\u6d41\u66a2\u985e\u7684\u5206\u985e\u6a21\u578b\u3002 \u53e5\u5b50\u4e4b\u9593\u7684\u7de8\u8f2f\u8ddd\u96e2\u7684\u65b9\u6cd5(Ji & Eisenstein, 2013; Cheng & Liang, 2005)\u3002\u7de8\u8f2f\u8ddd\u96e2\u901a\u5e38\u5b9a \u6211\u5011\u63d0\u51fa\u7684\u7279\u5fb5\u5247\u662f\u4fdd\u7559\u4e86\u8cc7\u6599\u7684\u500b\u6642\u9593\u9ede\u7684\u8a9e\u7fa9\u7279\u5fb5\u5206\u5e03\u60c5\u5f62\uff0c\u6211\u5011\u5c07\u9019\u6a23\u6709\u5404\u8cc7\u6599 \u7fa9\u70ba\u5c07\u6e90\u53e5\u7de8\u8f2f\u70ba\u76ee\u6a19\u53e5(\u5305\u62ec\u63d2\u5165\u3001\u522a\u9664\u6216\u5207\u63db)\u7684\u6210\u672c\u6b65\u9a5f\u3002\u6700\u5f8c\u4e00\u500b\u7279\u5fb5\u662f\u6bcf\u500b\u53e5 \u6574\u500b\u6642\u9593\u9ede\u8a9e\u7fa9\u5206\u5e03\u7684\u5411\u91cf\u4f5c\u70ba\u7279\u5fb5\uff0c\u8b93\u5206\u985e\u5668\u53bb\u5b78\u7fd2\u8a9e\u7fa9\u6642\u5e8f\u5206\u5e03\u8207\u6d41\u66a2\u5ea6\u7684\u95dc\u806f\u3002 \u5b50\u4e2d\u55ae\u8a5e\u91cd\u8907\u983b\u7387\u7684\u5e73\u5747\u503c\u3002 \u5716 1 \u70ba\u4ee5\u8a9e\u610f\u6d41\u66a2\u5ea6\u8868\u5fb5\u70ba\u7279\u5fb5\u4e4b\u5be6\u9a57\u67b6\u69cb\u5716\uff0c\u5be6\u9a57\u6240\u4f7f\u7528\u7684\u5206\u985e\u5668\u7686\u662f\u652f\u63f4\u5411\u91cf</td></tr><tr><td colspan=\"4\">\u6a5f(SVM)\u548c\u76f8\u95dc\u6027\u7684\u7279\u5fb5\u9078\u64c7\uff0c\u5dee\u5225\u50c5\u5728\u65bc\u4f7f\u7528\u4f86\u4f5c\u70ba\u8fa8\u8b58\u57fa\u6e96\u7684\u7279\u5fb5\u62bd\u63db\u6210\u4e0d\u540c\u7684\u5404 2.4 \u983b\u9006\u6587\u6a94\u983b\u7387(TF-IDF) \u9805\u7279\u5fb5\u3002\u7531\u65bc\u6bcf\u4e00\u500b\u5152\u7ae5\u6558\u4e8b\u5305\u62ec\u4e0d\u540c\u6578\u91cf\u7684\u8cc7\u6599\u6a23\u672c(\u53e5\u6578)\uff0c\u70ba\u4e86\u5c0d\u65bc\u55ae\u500b\u53d7\u8a66\u8005\u7d66 \u65b9\u6cd5\u4e3b\u8981\u7531\u8a5e\u983b(TF)\u4ee5\u53ca\u9006\u5411\u6587\u6a94\u983b\u7387(inverse document frequency\uff0cIDF)\u6240\u7d44\u6210\uff0c \u51fa\u4e00\u500b\u55ae\u4e00\u7684\u6a19\u7c64\uff0c\u6211\u5011\u6703\u5c0d\u9810\u6e2c\u9032\u884c\u6295\u7968\u4f86\u6c7a\u5b9a\u6700\u7d42\u500b\u4eba\u7684\u6a19\u7c64\u3002\u800c\u6e96\u78ba\u7387\u7684\u9a57\u8b49\u5247 \u5176\u6578\u5b78\u5f0f\u5b50\u8868\u9054\u6210: \u63a1\u7528\u4ea4\u53c9\u9a57\u8b49\uff0c\u4f7f\u7528\u7684\u8a55\u91cf\u65b9\u6cd5\u662f\u975e\u52a0\u6b0a\u5e73\u5747\u53ec\u56de\u7387(UAR)\u3002</td></tr><tr><td colspan=\"4\">TF , \u518d\u8005\uff0c\u6211\u5011\u4e5f\u5229\u7528\u6211\u5011\u63d0\u53d6\u51fa\u7684\u6d41\u66a2\u5ea6\u7279\u5fb5\u53bb\u5c0d\u81ea\u9589\u75c7\u89c0\u5bdf\u8a3a\u65b7\u91cf\u8868\u4e2d\u7684\u8a55\u5206\u7d30\u9805 , \u2211 , \u53bb\u505a\u9ad8\u4f4e\u5206\u5206\u985e\uff0c\u5176\u4e2d\u5305\u62ec\u523b\u677f\u7684\u4f7f\u7528\u55ae\u5b57\u6216\u7247\u8a9e\u7a0b\u5ea6\u3001\u6703\u8a71\u6d41\u66a2\u6027\uff0c\u4ee5\u53ca\u5831\u544a\u4e8b\u4ef6\u7684</td></tr><tr><td colspan=\"4\">b \u5206\u5b50\u662f\u7b2c i \u500b\u55ae\u8a5e\u5728\u7b2c j \u500b\u6587\u7ae0\u7684\u55ae\u8a5e\u8a08\u6578\uff0c\u5206\u6bcd\u662f\u7b2c i \u500b\u55ae\u8a5e\u5728\u5168\u90e8\u6587\u7ae0\u4e2d\u7684\u7e3d\u51fa\u73fe\u6b21 \u80fd\u529b\u3002</td></tr><tr><td>\u6578\u3002</td><td>\u2022</td><td>,</td><td>b</td></tr><tr><td/><td colspan=\"2\">\u2022 IDF \u503c\u8207\u76ee\u6a19\u8a5e\u5728\u6574\u500b\u8a9e\u6599\u5eab\u4e2d\u7684\u901a\u7528\u6027\u6709\u95dc\u3002 ,</td><td>b</td></tr><tr><td colspan=\"4\">\u5728\u7d30\u7bc0\u4e2d,\u5fd8\u8a18\u9598 ,\u9598\u53e3\u8ca0\u8cac\u63a7\u5236\u904e\u53bb\u7684\u8cc7\u8a0a\u88ab\u5141\u8a31\u901a\u904e\u4e3b\u8981\u7d30\u80de \u7684\u901a\u904e\u6bd4\u7387,\u6211\u5011\u5229 \u7528\u6b64\u4e00\u6a5f\u5236\u4f86\u7576\u4f5c\u5c0e\u51fa\u8cc7\u8a0a\u6d41\u66a2\u5ea6\u7684\u4e00\u7a2e\u6e2c\u91cf\u65b9\u6cd5\u3002\u6211\u5011\u9996\u5148\u4f7f\u7528\u76e3\u7763\u5f0f\u5b78\u7fd2\u9810\u8a13\u7df4\u5728 | | IDF 1 | \u2208 \u2236 \u2208 |</td></tr><tr><td colspan=\"4\">\u56db\u7a2e\u6545\u4e8b\u985e\u578b\u7684\u6a19\u7c64\u4e0a\uff0c\u4f5c\u70ba\u521d\u59cb\u5316\u7684\u6a21\u578b\u53c3\u6578\uff0c\u6b64\u9577\u77ed\u671f\u8a18\u61b6\u7db2\u7d61\u5c64\u7684\u96b1\u85cf\u7dad\u5ea6\u70ba 128 \u5206\u6bcd\u662f\u6574\u500b\u8a9e\u6599\u5eab\u7684\u6587\u7ae0\u6578\u91cf\uff0c\u5206\u6bcd\u662f\u76ee\u6a19\u8a5e\u51fa\u73fe\u7684\u6587\u7ae0\u6578\u91cf\u3002\u5982\u679c\u76ee\u6a19\u8a5e\u6c92\u6709\u51fa\u73fe\u5728 \u7dad\u3002 \u8a9e\u6599\u5eab\u4e2d\uff0c\u5206\u6bcd\u5c07\u70ba\u96f6\u3002\u5e38\u7528\u7684\u65b9\u6cd5\u662f\u8abf\u6574\u5206\u6bcd\u70ba1 | \u2208 \u2236 \u2208 |\u3002</td></tr><tr><td colspan=\"4\">2.2.5 \u8a9e\u610f\u6d41\u66a2\u5ea6\u8868\u5fb5 (Lexical Coherence Representation) \u6700\u7d42\u5f97\u5230\u503c\u7684\u5f0f\u5b50\u5982\u4e0b:</td></tr><tr><td colspan=\"4\">\u70ba\u4e86\u5f9e\u8cc7\u8a0a\u4e2d\u5c0e\u51fa\u6d41\u66a2\u5ea6\u7684\u8868\u9054\u7279\u5fb5\uff0c\u6211\u5011\u9996\u5148\u4f7f\u7528\u8cc7\u6599\u96c6\u4e8c\u56db\u985e\u7ae5\u8a71\u6545\u4e8b\uff0c\u8a13\u7df4\u51fa\u4f86 [Figure 2. Extraction procedure of cohesion matrix] \u4eab\u6a21\u578b\u53c3\u6578\uff0c\u53bb\u5c0d\u8cc7\u6599\u96c6\u4e00\u53bb\u9032\u884c\u9069\u6027\u5316\u7684\u5fae\u8abf\uff0c\u9019\u662f\u4e00\u7a2e\u8f49\u79fb\u5b78\u7fd2\u7684\u65b9\u5f0f\u3002\u7d93\u904e\u5c0d\u8cc7 \u5716 2. \u6d41\u66a2\u5ea6\u77e9\u9663\u7279\u5fb5\u64f7\u53d6\u7db2\u7ad9\u793a\u610f\u5716 \u7684\u6a21\u578b\u6b0a\u91cd\uff0c\u4f5c\u70ba\u6211\u5011\u9810\u8a13\u7df4\u597d\u7684\u521d\u59cb\u5316\u53c3\u6578\u3002\u7136\u5f8c\u6211\u5011\u7528\u6211\u5011\u9810\u8a13\u7df4\u904e\u7684\u6a21\u578b\u901a\u904e\u5206 t d t d t IDF TF IDF TF \uf0b4 \uf03d \uf02d , ,</td></tr></table>", | |
| "text": "", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| }, | |
| "TABREF3": { | |
| "content": "<table><tr><td>30</td><td>\u5289\u4e8e\u78a9 \u7b49 \u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u4e4b\u5fd8\u8a18\u9598\u63d0\u53d6\u8a9e\u610f\u6d41\u66a2\u5ea6\u4e4b\u67b6\u69cb\u4ee5\u81ea\u9589\u75c7\u5c0f\u5b69\u8aaa\u6545\u4e8b\u70ba\u4f8b 29 \u5289\u4e8e\u78a9 \u7b49</td></tr><tr><td colspan=\"2\">3.3 \u5be6\u9a57\u5206\u6790 (Analysis) \u8868 1Features Coh-Metrix TFIDF TFIDF + Coh-Metrix ** LSTM ** Lexical Coherence Representation ** \u5728\u5206\u985e\u6e96\u78ba\u5ea6\u65b9\u9762\uff0c\u8868 1 \u7e3d\u7d50\u6211\u5011\u7684\u5be6\u9a57\u7d50\u679c\u3002\u7372\u5f97\u7684\u6700\u4f73\u6e96\u78ba\u5ea6\u662f\u901a\u904e\u4f7f\u7528\u6211\u5011 UAR/p value Accuracy 0.73/0.12 0.74 0.77/0.10 0.78 0.80/0.04 0.79 0.85/0.03 0.86 0.92/0.007 0.91 \u63d0\u51fa\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u8868\u5fb5\uff0c\u9054\u5230\u4e86 UAR 92%\u3002\u5b83\u6bd4\u5176\u4ed6\u6bd4\u8f03\u7684\u7279\u5fb5\u6d41\u66a2\u5ea6\u77e9\u9663\u548c TFIDF \u65b9\u6cd5\u7684 73%\u548c 77%\u90fd\u8981\u51fa\u8272\uff0c\u4e14\u662f\u986f\u8457\u4f86\u7684\u9ad8\u3002\u800c\u4e14\uff0c\u6709\u8da3\u7684\u4e00\u9ede\u662f\u7576\u8207\u4f7f\u7528\u9577\u77ed\u671f\u8a18 \u61b6\u6a21\u578b\u76f4\u63a5\u57f7\u884c\u5206\u985e\u7684\u7279\u5fb5\u64f7\u53d6\uff0c\u53bb\u9032\u884c\u6bd4\u8f03\u6642\uff0c\u4f7f\u7528\u6211\u5011\u6240\u63d0\u51fa\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u6240 \u8a13\u7df4\u7684\u6a21\u578b\uff0c\u5728\u8fa8\u8b58\u5178\u578b\u5b69\u7ae5\u8207\u81ea\u9589\u75c7\u5b69\u7ae5\u6587\u672c\u5167\u5bb9\u7684\u4efb\u52d9\u4e0a\u9084\u80fd\u6709 8%\u7684\u53ec\u56de\u7387\u512a\u52e2\uff0c \u6211\u5011\u6240\u63d0\u51fa\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u662f\u7531\u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u7684\u5167\u90e8\u53c3\u6578\u884d\u751f\u800c\u4f86\u7684\uff0c\u800c\u5176\u4f3c\u4e4e\u6bd4 \u4f7f\u7528\u6574\u500b\u9577\u77ed\u671f\u7279\u5fb5\u6a21\u578b\u63d0\u53d6\u7684\u7279\u5fb5\u66f4\u5177\u6709\u6307\u6a19\u6027\u3002\u6700\u5f8c\uff0c\u5c0d\u65bc\u4e0d\u540c\u53e5\u5b50\u6578\u91cf\u4f86\u505a\u8cc7\u6599 \u64f4\u5145\u5c0d\u65bc\u6e96\u78ba\u7387\u7684\u5f71\u97ff\uff0c\u8868 2 \u5217\u51fa\u4e86\u5be6\u9a57\u7684\u7d50\u679c\u3002\u8cc7\u6599\u64f4\u5145\u6b65\u9a5f\u4e2d\u9078\u5b9a\u7684\u4f7f\u7528\u53e5\u6578\uff0c\u6700 \u4f73\u6578\u91cf\u4f3c\u4e4e\u5728 n=5 \u5de6\u53f3\u6642\u6703\u5f97\u5230\u6700\u597d\u7684\u6548\u679c\u3002 \u8868 2. \u4e0d\u540c\u8cc7\u6599\u64f4\u589e\u53c3\u6578\u4e0b\u7684\u6e96\u78ba\u7387 [Table 2. Performance of different augment set] # of sentence LSTM Lexical Coherence n=1 0.76 0.78 n=3 0.85 0.87 n=5 0.78 0.92 n=7 0.79 0.81 n=9 0.71 0.81 \u800c\u5728\u5c0d\u81ea\u9589\u75c7\u8a3a\u65b7\u89c0\u5bdf\u91cf\u8868(ADOS)\u6709\u95dc\u53e3\u8a9e\u80fd\u529b\u7684\u8a55\u5206\u7d30\u9805\u4e2d(\u8868 3)\uff0c\u6211\u5011\u767c\u73fe\u6b64 \u7bc7\u6240\u63d0\u51fa\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\uff0c\u7684\u78ba\u8207\u8a55\u5206\u7d30\u9805\u4e2d\u8ddf\u6d41\u66a2\u5ea6\u6700\u76f8\u95dc\u7684\"\u6703\u8a71\u6d41\u66a2\u6027\"\u6700\u76f8 \u95dc\uff0c\u5728\u9ad8\u4f4e\u5206\u5169\u985e\u5206\u985e\u7684\u4efb\u52d9\u4e2d\u80fd\u9054\u5230 80.48%\u7684\u6e96\u78ba\u7387\u3002 \u8868 3. \u6d41\u66a2\u5ea6\u76f8\u95dc\u884c\u70ba\u5c6c\u6027 [Table 3. Attributes of behavior related to coherence] \u7de8\u865f \u7d30\u9805\u5167\u5bb9 \u9ad8\u4f4e\u5206\u985e\u6e96\u78ba\u7387 A4 \u523b\u677f\u7684\u4f7f\u7528\u55ae\u5b57\u6216\u7247\u8a9e 0.5378 A8 \u6703\u8a71\u6d41\u66a2 0.8048 A7 \u5831\u544a\u4e8b\u4ef6 0.4208 \u5716 3. \u7531\u5de6\u81f3\u53f3\u5206\u5225\u70ba\u6253\u4e82\u5b57\u5e8f\u3001\u53e5\u5b50\u9806\u5e8f\u3001\u522a\u9664\u9023\u8cab\u53e5\u5b50\u4e2d\u7279\u5b9a\u5b57\u8a5e [Figure 3. From left to right, the green dots represent the samples of random word order, random sentence order, delete random word in sentences] \u6211\u5011\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u6a21\u578b\u5be6\u73fe\u4e86\u9ad8\u8b58\u5225\u6e96\u78ba\u6027\uff0c\u4f46\u7531\u65bc\u6a21\u578b\u7684\u8907\u96dc\u6027\uff0c\u5b83\u7f3a\u4e4f\u4e86\u7279\u5fb5 \u7684\u76f4\u63a5\u89e3\u91cb\u6027\u3002\u56e0\u6b64\uff0c\u6211\u5011\u63a1\u7528\u4e86\u985e\u6700\u8fd1\u767c\u8868\u65bc ICLR \u7684\u985e\u4f3c\u65b9\u6cd5(Li et al., 2016; Koh & Liang, 2017)\u4f86\u9032\u4e00\u6b65\u77ad\u89e3\u5206\u6790\u6211\u5011\u7684\u7279\u5fb5\u3002\u5728\u9019\u500b\u5206\u6790\u4e2d\uff0c\u6211\u5011\u9996\u5148\u901a\u904e\u96a8\u6a5f\u6253\u4e82\u5178\u578b \u5c0f\u5b69\u8cc7\u6599\u6a23\u672c\u4e2d\u7684\u55ae\u8a5e\u9806\u5e8f\u3001\u53e5\u5b50\u9806\u5e8f\uff0c\u6216\u662f\u522a\u9664\u9023\u8cab\u53e5\u4e2d\u7684\u96a8\u6a5f\u8a5e\u8a9e\uff0c\u6a21\u64ec\u51fa\u8a9e\u7121\u502b \u6b21\u7684\u6a23\u672c\uff0c\u5229\u7528\u6f5b\u5728\u8a9e\u7fa9\u5206\u6790\u4f86\u964d\u7dad\u5c07\u7279\u5fb5\u8868\u793a\u6210\u4e8c\u7dad\u5716\uff0c\u4f86\u89c0\u5bdf\u8996\u89ba\u5316\u8cc7\u6599\u5206\u5e03\u3002\u7136 \u5f8c\u6211\u5011\u9032\u4e00\u6b65\u540c\u6642\u4f7f\u7528\u5169\u7a2e\u6211\u5011\u4e0a\u8ff0\u6240\u6a21\u64ec\u7684\u4e0d\u6d41\u66a2\u6a5f\u5236(\u540c\u6642\u6253\u4e82\u5b57\u5e8f\u6216\u53e5\u5e8f\u3001\u540c\u6642\u522a \u9664\u7279\u5b9a\u5b57\u8207\u6253\u4e82\u53e5\u5e8f)\uff0c\u4f86\u89c0\u5bdf\u4e0d\u540c\u7a0b\u5ea6\u7684\u4e0d\u6d41\u66a2\u60c5\u5f62\u5728\u6211\u5011\u7684\u7a7a\u9593\u4e2d\u7684\u8cc7\u6599\u5206\u5e03\u72c0\u6cc1\u3002 \u5716 4. \u7531\u5de6\u81f3\u53f3\u5206\u5225\u70ba\u540c\u6642\u6253\u4e82\u5b57\u5e8f\u6216\u53e5\u5e8f\u3001\u540c\u6642\u522a\u9664\u7279\u5b9a\u5b57\u8207\u6253\u4e82\u53e5\u5e8f [Figure 4. From left to right, the green dots represent the samples of random word and sentence order, delete random word and random sentence order] \u85cd\u9ede\u8868\u793a\u5178\u578b\u5b69\u7ae5\u7684\u8cc7\u6599\u6a23\u672c\uff0c\u7d05\u9ede\u4ee3\u8868\u81ea\u9589\u75c7\u5b69\u7ae5\u7684\u8cc7\u6599\u6a23\u672c\uff0c\u9ec3\u9ede\u4ee3\u8868\u5178\u578b\u5b69 \u7ae5\u6545\u4e8b\u7684\u6a23\u672c\u901a\u904e\u6a21\u64ec\u6558\u8ff0\u4e0d\u6d41\u66a2\u7684\u6a5f\u5236\u6539\u8b8a\u5f8c\u7684\u8cc7\u6599\u5206\u5e03\u3002\u89c0\u5bdf\u5716 4 \u80fd\u6709\u8da3\u7684\u767c\u73fe\uff0c \u9ec3\u8272\u7684\u9ede\u5728\u5178\u578b\u5b69\u7ae5\u8207\u81ea\u9589\u75c7\u5b69\u7ae5\u4e4b\u9593\uff0c\u986f\u793a\u6211\u5011\u7684\u8868\u5fb5\u80fd\u53cd\u6620\u51fa\u6211\u5011\u4e0a\u8ff0\u4e09\u7a2e\u4e0d\u6d41\u66a2 \u6a5f\u5236\u7684\u60c5\u6cc1\u3002\u6b64\u5916\uff0c\u96a8\u8457\u6211\u5011\u5f15\u5165\u66f4\u591a\u7684\u4e0d\u6d41\u66a2\u6a5f\u5236(\u540c\u6642\u6253\u4e82\u5b57\u5e8f\u6216\u53e5\u5e8f\u3001\u540c\u6642\u522a\u9664\u7279 \u5b9a\u5b57\u8207\u6253\u4e82\u53e5\u5e8f)\u9032\u5165\u5178\u578b\u5b69\u7ae5\u6a23\u672c\uff0c\u9ec3\u9ede\u6703\u66f4\u9760\u8fd1\u4e0d\u6d41\u66a2\u7684\u81ea\u9589\u75c7\u6a23\u672c\u3002\u9019\u500b\u5be6\u9a57\u5206\u6790 \u8868\u660e\uff0c\u6211\u5011\u63d0\u51fa\u7684\u7279\u5fb5\u80fd\u5920\u53cd\u6620\u4e0d\u540c\u7684\u6d41\u66a2\u6a5f\u5236\uff0c\u4e26\u4e14\u8861\u91cf\u6587\u672c\u4e0d\u540c\u7684\u6d41\u66a2\u7a0b\u5ea6\u3002 \u5716 5. \u6a23\u672c\u503c\u5206\u5e03\u5716 [Figure 5. histogram of samples] \u6211\u5011\u4e5f\u756b\u51fa\u81ea\u9589\u75c7\u6a23\u672c\u3001\u5178\u578b\u5b69\u7ae5\u6a23\u672c\uff0c\u4ee5\u53ca\u5c0e\u5165\u4e0d\u6d41\u66a2\u6a5f\u5236\u7684\u5178\u578b\u5b69\u7ae5\u6a23\u672c\u7684\u503c \u5206\u5e03\u5716\uff0c\u5f9e\u5716 5 \u80fd\u767c\u73fe\uff0c\u5c0e\u5165\u4e0d\u6d41\u66a2\u6a5f\u5236\u7684\u5178\u578b\u5b69\u7ae5\u6a23\u672c\uff0c\u7684\u78ba\u5728\u8cc7\u6599\u5206\u5e03\u4e0a\u66f4\u8da8\u8fd1\u65bc \u4e0d\u6d41\u66a2\u7684\u81ea\u9589\u75c7\u5b69\u7ae5\u6a23\u672c\u3002\u9032\u4e00\u6b65\u6211\u5011\u4e5f\u4ed4\u7d30\u6bd4\u8f03\u5404\u6a23\u672c\u4e2d\u7684\u7d71\u8a08\u503c\uff0c\u4e26\u756b\u6210\u6a23\u672c\u503c\u5206 \u5e03\u5716\u4f86\u89c0\u5bdf\u8cc7\u6599\u5206\u5e03\u7684\u60c5\u6cc1\uff0c\u5728\u6700\u5927\u503c(Max)\u3001\u504f\u5ea6\u7684\u6a23\u672c\u503c\u5206\u5e03\u5716\u4e2d\uff0c\u6211\u5011\u80fd\u660e\u986f\u770b\u51fa \u9ec3\u8272\u7684\u8cc7\u6599\u5206\u5e03\u5f9e\u539f\u85cd\u8272\u5206\u5e03\u5340\u57df\u79fb\u5f80\u7d05\u8272\u7684\u8cc7\u6599\u5206\u5e03\u5340\u57df\u3002 \u7d93\u7db2\u7d61\u67b6\u69cb\u5167\u7684\u907a\u5fd8\u9598\u4e2d\uff0c\u8cc7\u8a0a\u96a8\u6642\u9593\u50b3\u5c0e\u7684\u6a5f\u5236\uff0c\u6a21\u64ec\u4eba\u985e\u95b1\u8b80\u6642\u6703\u5c0d\u4e0d\u540c\u6642\u9593\u9ede\u8cc7 \u8a0a\u6709\u4e0d\u540c\u6b0a\u91cd\u7684\u7279\u6027\uff0c\u4f86\u63a8\u7b97\u51fa\u6d41\u66a2\u5ea6\u7279\u5fb5\u3002\u7528\u6211\u5011\u6240\u63d0\u51fa\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u8a13\u7df4\u7684\u6a21\u578b\uff0c \u5728\u81ea\u9589\u75c7\u8fa8\u8b58\u80fd\u9054\u5230 92%\u7684\u8b58\u5225\u6e96\u78ba\u7387\uff0c\u8b49\u660e\u512a\u65bc\u50b3\u7d71\u8861\u91cf\u8a5e\u5f59\u5167\u5bb9\u548c\u9023\u8cab\u6027\u7684\u65b9\u6cd5\u3002 \u6700\u5f8c,\u900f\u904e\u5728\u5178\u578b\u5b69\u7ae5\u7684\u8cc7\u6599\u6a23\u672c\u4e2d\u5c0e\u5165\u985e\u4f3c\u4e0d\u6d41\u66a2\u7684\u7279\u5fb5(\u985e\u6bd4\u8a9e\u7121\u502b\u6b21\u7684\u60c5\u6cc1)\uff0c\u6211\u5011 \u53ef\u4ee5\u5f9e\u8996\u89ba\u5316\u5f8c\u7684\u8cc7\u6599\u6a23\u672c\u5206\u4f48\u9ede\u5716\u89c0\u5bdf\u5230\uff0c\u5c07\u8cc7\u6599\u6a23\u672c\u6620\u5c04\u5230\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u7684\u7a7a\u9593 \u5f8c\uff0c\u88ab\u5c0e\u5165\u4e0d\u6d41\u66a2\u7279\u5fb5\u7684\u5178\u578b\u5c0f\u5b69\u6a23\u672c\u9ede\u6703\u79fb\u52d5\u5f97\u66f4\u52a0\u63a5\u8fd1\u65bc\u81ea\u9589\u75c7\u5c0f\u5b69\u6a23\u672c\u3002\u9019\u6a23\u7684 \u7d50\u679c\u63a8\u8ad6\uff0c\u5373\u4f7f\u6211\u5011\u7684\u9577\u77ed\u671f\u8a18\u61b6\u6a21\u578b\u67b6\u69cb\u4e26\u975e\u76f4\u63a5\u5c0d\u6d41\u66a2\u5ea6\u6a19\u7c64\u505a\u5b78\u7fd2\uff0c\u4f46\u5728\u5176\u5167\u90e8 \u67b6\u69cb\u7684\u5143\u7d20\u4e2d\u4f3c\u4e4e\u542b\u6709\u4ee3\u8868\u8a9e\u610f\u9023\u8cab\u6027\u7684\u6d41\u66a2\u5ea6\u7684\u5143\u7d20\uff0c\u751a\u81f3\u9084\u80fd\u5be6\u73fe\u8f03\u9ad8\u7684\u6e96\u78ba\u7387\u3002 \u800c\u5728\u672a\u4f86\u767c\u5c55\u65b9\u5411\uff0c\u9996\u5148\u662f\u80fd\u53bb\u5206\u6790\u6211\u5011\u6240\u5c0e\u51fa\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u8207\u8a08\u7b97\u8a9e\u8a00\u5b78\u4e2d\u8a2d\u8a08 \u7684\u7279\u5fb5\u4e4b\u9593\u7684\u95dc\u4fc2\uff0c\u82e5\u80fd\u627e\u5230\u8207\u8a08\u7b97\u8a9e\u8a00\u5b78\u4e2d\u8a2d\u8a08\u7684\u7279\u5fb5\u5982\u8a9e\u610f\u3001\u95dc\u9375\u5b57\u3001\u8a9e\u6cd5\u91cd\u8907\u9019 \u4e9b\u7279\u5fb5\u4e4b\u9593\u7684\u95dc\u4fc2\uff0c\u4fbf\u80fd\u66f4\u52a0\u77ad\u89e3\u795e\u7d93\u7db2\u7d61\u7406\u89e3\u6587\u7ae0\u7684\u65b9\u5f0f\u3002\u6b64\u5916\uff0c\u6d41\u66a2\u5ea6\u4e5f\u53ef\u4ee5\u5f80\u591a \u6a21\u614b\u7684\u65b9\u5411\u767c\u5c55\uff0c\u5982\u8a9e\u8abf\u6d41\u66a2\u548c\u624b\u52e2\u5354\u8abf\u6216\u81c9\u90e8\u3001\u80a2\u9ad4\u52d5\u4f5c\u65b9\u9762\u3002\u6700\u5f8c\u672a\u4f86\u53ef\u4ee5\u5957\u7528\u66f4 \u65b0\u7684\u67b6\u69cb\u4f8b\u5982 Attention-based LSTM\uff0c\u4e26\u61c9\u7528\u5728\u73fe\u5be6\u4e16\u754c\u7684\u884c\u70ba\u8cc7\u8a0a\u5206\u6790\uff0c\u63d0\u9ad8\u81e8\u5e8a\u50f9\u503c \u5c07\u7e7c\u7e8c\u662f\u7814\u7a76\u7684\u4e2d\u5fc3\u76ee\u6a19\u3002 \u5716 6\u672c\u6587\u63d0\u51fa\u4e86\u4e00\u7a2e\u65b0\u7684\u8cc7\u6599\u5c0e\u5411\u7684\u8a9e\u610f\u6d41\u66a2\u5ea6\u7279\u5fb5\u5b78\u7fd2\u67b6\u69cb\uff0c\u5176\u7cbe\u96a8\u662f\u5229\u7528\u9577\u77ed\u671f\u8a18\u61b6\u795e \u53c3\u8003\u6587\u737b (References)</td></tr></table>", | |
| "text": "", | |
| "html": null, | |
| "num": null, | |
| "type_str": "table" | |
| } | |
| } | |
| } | |
| } |