{ "paper_id": "O09-5001", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T08:11:15.723825Z" }, "title": "Modeling Taiwanese POS Tagging Using Statistical Methods and Mandarin Training Data", "authors": [ { "first": "Un-Gian", "middle": [], "last": "Iunn", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Taiwan University", "location": { "settlement": "Taipei", "country": "Taiwan" } }, "email": "" }, { "first": "Jia-Hung", "middle": [], "last": "Tai", "suffix": "", "affiliation": {}, "email": "" }, { "first": "Kiat-Gak", "middle": [], "last": "Lau", "suffix": "", "affiliation": {}, "email": "kiatgak01@gmail.com" }, { "first": "Cheng-Yan", "middle": [], "last": "Kao", "suffix": "", "affiliation": { "laboratory": "", "institution": "National Taiwan University", "location": { "settlement": "Taipei", "country": "Taiwan" } }, "email": "cykao@csie.ntu.edu.tw" }, { "first": "Keh-Jiann", "middle": [], "last": "Chen", "suffix": "", "affiliation": {}, "email": "kchen@iis.sinica.edu.tw" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "In this paper, we introduce a POS tagging method for Taiwan Southern Min. We use the more than 62,000 entries of the Taiwanese-Mandarin dictionary and 10 million words of Mandarin training data to tag Taiwanese. The literary written Taiwanese corpora have both Romanized script and Han-Romanization mixed script, and include prose, novels, and dramas. We follow the tagset drawn up by CKIP. We developed a word alignment checker to assist with the word alignment for the two scripts. It searches the Taiwanese-Mandarin dictionary to find corresponding Mandarin candidate words, selects the most suitable Mandarin word using an HMM probabilistic model from the Mandarin training data, and tags the word using an MEMM classifier. We achieve an accuracy rate of 91.6% on Taiwanese POS tagging work, and we analyze the errors. We also discover some preliminary Taiwanese training data.", "pdf_parse": { "paper_id": "O09-5001", "_pdf_hash": "", "abstract": [ { "text": "In this paper, we introduce a POS tagging method for Taiwan Southern Min. We use the more than 62,000 entries of the Taiwanese-Mandarin dictionary and 10 million words of Mandarin training data to tag Taiwanese. The literary written Taiwanese corpora have both Romanized script and Han-Romanization mixed script, and include prose, novels, and dramas. We follow the tagset drawn up by CKIP. We developed a word alignment checker to assist with the word alignment for the two scripts. It searches the Taiwanese-Mandarin dictionary to find corresponding Mandarin candidate words, selects the most suitable Mandarin word using an HMM probabilistic model from the Mandarin training data, and tags the word using an MEMM classifier. We achieve an accuracy rate of 91.6% on Taiwanese POS tagging work, and we analyze the errors. We also discover some preliminary Taiwanese training data.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Abstract", "sec_num": null } ], "body_text": [ { "text": "There are about 46 million Southern Min speakers in the world. If we list languages by the size of their speaking population, Southern Min is ranked 21. The Southern Min speakers are mainly distributed in eight countries (Gordon, 2005) . It is an important language that has received very little attention.", "cite_spans": [ { "start": 221, "end": 235, "text": "(Gordon, 2005)", "ref_id": "BIBREF7" } ], "ref_spans": [], "eq_spans": [], "section": "Background", "sec_num": "1.1" }, { "text": "The percentage of Southern Min speakers in Taiwan was over 70% (Huang, 1995) . Taiwan has the highest percentage of Southern Min speakers in the world. We will call this language as \"Taiwanese\" for simplification in this paper.", "cite_spans": [ { "start": 63, "end": 76, "text": "(Huang, 1995)", "ref_id": "BIBREF8" } ], "ref_spans": [], "eq_spans": [], "section": "Background", "sec_num": "1.1" }, { "text": "Many different types of written Taiwanese systems exist. Among these systems, the Han character script and one of the Romanized scripts (Pe\u030d h-\u014de-j\u012b, \u767d\u8a71\u5b57, abbrev. POJ, vernacular writing) are the most popular. Also, the mixture of the above two scripts, called the Han-Romanization mixed script (abbrev. as HR mixed script), has been adopted by many people .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Background", "sec_num": "1.1" }, { "text": "In order to establish the bases of written Taiwanese processing, we have constructed some tools over the past few years, including an online Taiwanese syllable dictionary (Iunn, 2003a) ; an online Taiwanese-Mandarin dictionary (abbrev. OTMD) (Iunn, 2000 (Iunn, , 2003b ; a 5,800,000 syllable HR mixed script and 3,400,000 syllable POJ script Taiwanese corpus; the online Taiwanese concordancer system based on this corpus (Iunn, 2003c; ; preliminary Taiwanese word frequency reports for the Taiwanese POJ and HR mixed scripts, based on the above Taiwanese corpus (Iunn, 2005) ; the digital archive database for written Taiwanese (abbrev. DADWT) literature data with POJ and HR mixed script paragraph alignment (Iunn, 2007) ; etc.", "cite_spans": [ { "start": 171, "end": 184, "text": "(Iunn, 2003a)", "ref_id": "BIBREF10" }, { "start": 242, "end": 253, "text": "(Iunn, 2000", "ref_id": "BIBREF9" }, { "start": 254, "end": 268, "text": "(Iunn, , 2003b", "ref_id": "BIBREF11" }, { "start": 422, "end": 435, "text": "(Iunn, 2003c;", "ref_id": "BIBREF12" }, { "start": 563, "end": 575, "text": "(Iunn, 2005)", "ref_id": "BIBREF13" }, { "start": 710, "end": 722, "text": "(Iunn, 2007)", "ref_id": "BIBREF14" } ], "ref_spans": [], "eq_spans": [], "section": "Motivation", "sec_num": "1.2" }, { "text": "We intend to annotate the Taiwanese corpus with POS markers for more advanced applications, including Taiwanese tone sandhi TTS system improvement , Taiwanese Treebank construction, etc.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Motivation", "sec_num": "1.2" }, { "text": "The primary difficulty encountered in the POS tagging of Taiwanese corpora is the question, \"What is the Taiwanese POS tagset?\" To date, no standard tagset for Taiwanese has been proposed. Under the circumstances, we have temporarily employed the Chinese POS tagset established by the CKIP Group of Academia Sinica (CKIP, 1993) . Unfortunately, we still encountered some problems because we did not have a Taiwanese dictionary that contained", "cite_spans": [ { "start": 315, "end": 327, "text": "(CKIP, 1993)", "ref_id": "BIBREF3" } ], "ref_spans": [], "eq_spans": [], "section": "Problem", "sec_num": "1.3" }, { "text": "the Mandarin POS tagset. The existing Taiwanese dictionaries merely contain basic vocabulary words, that is, nouns, verbs, adjectives, etc.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Using Statistical Methods and Mandarin Training Data", "sec_num": null }, { "text": "Moreover, there was another problem to surmount -manpower shortage. We did not have enough manpower to fully execute the POS tagging of the Taiwanese corpora.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Using Statistical Methods and Mandarin Training Data", "sec_num": null }, { "text": "Therefore, we proposed employing statistical procedures with the existing Mandarin resources and the OTMD to automatically complete the Taiwanese POS tagging. We used the Mandarin language model under the assumption that the word sequence in Taiwanese is similar to Mandarin. Shi (2006) translated the Mandarin sentences in the book, \"Modern Chinese 800 words '\u73fe\u4ee3 \u6f22\u8a9e\u516b\u767e\u8a5e' \" (by Shu-xiang L\u00fc) into Taiwanese and Hakka to establish the T3 corpus and developed some editing tools to help in the construction of the T3 Treebank. Chou (2006) used the Brill tagger based on the HMM model to tag words in the T3 Treebank. They used a tagset size of 26, and attained tagging accuracy rates of 92.80% and 85.59% for the training and test data, respectively. T3 Treebank has not been released publicly. Thus, we decided to use different tagsets and different tagged corpora in our experiments. Step 1", "cite_spans": [ { "start": 276, "end": 286, "text": "Shi (2006)", "ref_id": "BIBREF26" }, { "start": 524, "end": 535, "text": "Chou (2006)", "ref_id": "BIBREF2" } ], "ref_spans": [], "eq_spans": [], "section": "Using Statistical Methods and Mandarin Training Data", "sec_num": null }, { "text": "Step 2", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "POS Tagging Method", "sec_num": "2." }, { "text": "Step 3", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "POS Tagging Method", "sec_num": "2." }, { "text": "Step 4", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "POS Tagging Method", "sec_num": "2." }, { "text": "At first, the text contains both POJ and HR mixed scripts with paragraph by paragraph alignment.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "Step 1 converts the texts to word alignment form.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "Step 2 adds the Mandarin candidate words (translations).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "Step 3 selects the best Mandarin translation using the HMM model. Finally, we decide the POS tagging of each word using the MEMM model. The following subsection will describe this process in detail.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "For example, the original texts are \"T\u00e2i-\u00f4an t\u0113-it k\u00f4an \u00ea Gio\u030d k-san \u00ea h\u016b-k\u016bn khah k\u0113 \u00ea s\u00f3 . -ch\u0101i ... \" and", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "\"\u53f0\u7063 \u7b2c\u4e00 \u61f8 \u00ea \u7389\u5c71 \u00ea \u9644\u8fd1 \u8f03 \u4f4e \u00ea \u6240\u5728 ... \"", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "Taiwan first high of Mt.Jade of nearby more low of place", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "Step 1 converts the texts to word alignment form:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "\"\u53f0\u7063/T\u00e2i-\u00f4an \u7b2c\u4e00/t\u0113-it \u61f8/k\u00f4an \u00ea/\u00ea \u7389\u5c71/Gio\u030d k-san \u00ea/\u00ea \u9644\u8fd1/h\u016b-k\u016bn \u8f03/khah \u4f4e/k\u0113 \u00ea/\u00ea \u6240\u5728/s\u00f3 . -ch\u0101i \u2026 \" Then,", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "Step 2 adds the Mandarin translations:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "\"\u53f0\u7063/T\u00e2i-\u00f4an{\u53f0\u7063} \u7b2c\u4e00/t\u0113-it{\u7b2c\u4e00;\u7d55\u9802} \u61f8/k\u00f4an{\u9ad8} \u00ea/\u00ea{\u7684} \u7389\u5c71 /Gio\u030d k-san{\u7389\u5c71} \u00ea/\u00ea{\u7684} \u9644\u8fd1/h\u016b-k\u016bn{\u9644\u8fd1} \u8f03/khah{\u8f03} \u4f4e/k\u0113{\u4f4e} \u00ea/\u00ea{\u7684} \u6240\u5728/s\u00f3 . -ch\u0101i(\u53bb \u8655;", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "\u5730\u65b9;\u89d2\u982d;\u6240\u5728;\u8655\u6240;\u5834\u6240;\u9593\uf97e} \u2026\"", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "Step 3 selects the best Mandarin translation using the HMM model (we omit the original Taiwanese texts):", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "\"\u53f0\u7063 \u7b2c\u4e00 \u9ad8 \u7684 \u7389\u5c71 \u7684 \u9644\u8fd1 \u8f03 \u4f4e \u7684 \u5730\u65b9 \u2026\" Finally,", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "Step 4 decides the POS tagging of each word using the MEMM model:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "\"\u53f0\u7063/T\u00e2i-\u00f4an(Nc) \u7b2c\u4e00/t\u0113-it(Neu) \u61f8/k\u00f4an(VH) \u00ea/\u00ea(DE) \u7389\u5c71/ Gio\u030d k -san(Nc) \u00ea/\u00ea(DE) \u9644\u8fd1/h\u016b-k\u016bn(Nc) \u8f03/khah(Dfa) \u4f4e/k\u0113(VH) \u00ea/\u00ea(DE) \u6240\u5728/s\u00f3 . -ch\u0101i(Na)\u2026 \"", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "We will illustrate our work with Figure 1 in the following subsections.", "cite_spans": [], "ref_spans": [ { "start": 33, "end": 41, "text": "Figure 1", "ref_id": null } ], "eq_spans": [], "section": "Figure 1. Taiwanese POS Tagging System Architecture Diagram", "sec_num": null }, { "text": "The corpus we chose is part of the DADWT project achievements of the National Museum of Taiwan Literature. It contains both POJ and HR mixed scripts with paragraph by paragraph alignment, including novels, prose, dramas, and poems (Iunn, 2007) .", "cite_spans": [ { "start": 231, "end": 243, "text": "(Iunn, 2007)", "ref_id": "BIBREF14" } ], "ref_spans": [], "eq_spans": [], "section": "Origin of the Corpus", "sec_num": "2.1" }, { "text": "First, we developed a word alignment program to aid manual processing. We arranged the word alignment of the two scripts, where the paragraphs were already aligned. This program not only collates the number of syllables in the two scripts, but it also compares and contrasts the two scripts with the entries of the OTMD. If the program does not find the two scripts within the same entry, it highlights the corresponding words to remind the user that the word may be an unknown word, an inconsistent usage of the Han character, or a typographical error.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Word by Word Alignment", "sec_num": "2.2" }, { "text": "The OTMD was announced and has been online since 2000. The main data provider is Robert L. Cheng, but many anonymous contributors also offer entries and correct the typographical errors. There are a total of more than 62,000 entries. The URL is http://iug.csie.dahan.edu. tw/q. This dictionary offers POJ, HR mixed script, and Mandarin fields, with the POJ field also offering the different accents. The pronunciation function was added in 2006, and English translation was added to more than 10,000 entries in 2007 based on Embree (1984) , which contains English, Mandarin, and POJ fields.", "cite_spans": [ { "start": 525, "end": 538, "text": "Embree (1984)", "ref_id": "BIBREF5" } ], "ref_spans": [], "eq_spans": [], "section": "Word by Word Alignment", "sec_num": "2.2" }, { "text": "Next, we continued to search for the corresponding Mandarin candidate words from the POJ and HR mixed script word pairs via the OTMD. The mapping was one-to-many. In short, a Taiwanese word pair would have more than one Mandarin word counterpart. For example, \"\u611b /\u00e0i\" in Taiwanese has the meanings of \"\u611b\"'love (person),' \"\u559c\u6b61\"'like (thing),' \"\u8981\" 'want to,' \"\u9700\u8981\"'need to,' etc. in Mandarin. Nevertheless, we were not able to find counterparts for certain words, since they were not contained in the OTMD. We also found some that had different HR mixed script usage.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Finding the Corresponding Mandarin Candidate Words", "sec_num": "2.3" }, { "text": "For instance, the Taiwanese word that appears as \"\u8f03\u8d0f/khah-i\u00e2\u207f\" 'more than' in the corpus appears as \"khah \u8d0f/khah-i\u00e2\u207f\" in the dictionary. With regard to problems of this nature, we applied the following solution. If the POJ and HR mixed script word pair could not be found, we temporarily removed the HR mixed script and searched for the Mandarin word counterpart again using the POJ script. If the characters of HR mixed script were all Han characters, we regarded the Han characters as one of a Mandarin candidate word (assuming that the word is common to both Taiwanese and Mandarin).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Finding the Corresponding Mandarin Candidate Words", "sec_num": "2.3" }, { "text": "This method might increase the number of the Mandarin candidate words, especially for single syllable words. For instance, the word pair \"\u8f49/ch\u014dan\"'turn' appears in the text. We could not find an entry that contains both \"\u8f49\" and \"ch\u014dan\" in the OTMD. The corresponding Mandarin translations of \"ch\u014dan\" in the dictionary are \"\u626d\"'twist' and \"\u4e0a\" 'up'. We added \"\u8f49\"'turn' as the supplementary Mandarin translation, but the meanings of these three words differ. Note: There exists not \"\u8f49/ch\u014dan\" entry in the OTMD. The Mandarin translation of \"\u8f49/ch\u014dan\" will be \"\u626d,\" \"\u4e0a\" and \"\u8f49\"", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Finding the Corresponding Mandarin Candidate Words", "sec_num": "2.3" }, { "text": "If the strategy was still unable to find any results, the HR mixed script was directly recognized as the Mandarin candidate word. For instance, no dictionary entry was found for the word pair appearing as \"\u6709\u5f62/i\u00fa-h\u00eang\"'tangible' in the text, neither could one be found in the search using the POJ script \"i\u00fa-h\u00eang.\" So, the HR mixed script \"\u6709\u5f62\" was directly recognized as the Mandarin candidate word (Lau, 2007) .", "cite_spans": [ { "start": 398, "end": 409, "text": "(Lau, 2007)", "ref_id": "BIBREF19" } ], "ref_spans": [], "eq_spans": [], "section": "Finding the Corresponding Mandarin Candidate Words", "sec_num": "2.3" }, { "text": "We employed the Hidden Markov Model and Viterbi algorithm, and we made use of the bigram word training data of the ten-million word balanced Sinica corpus of the CKIP Group of Academia Sinica to select the most appropriate corresponding Mandarin word from the Mandarin candidate words. Figure 2 is an example. The selected words are boxed and bold.", "cite_spans": [], "ref_spans": [ { "start": 286, "end": 294, "text": "Figure 2", "ref_id": null } ], "eq_spans": [], "section": "Selecting the Best Mandarin Translation", "sec_num": "2.4" }, { "text": "Word", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Taiwanese", "sec_num": null }, { "text": "\u5c0d/ T\u00f9i \u53e4\u65e9/ k\u00f3 . -ch\u00e1 \u4ee5\uf92d/ \u00ed-l\u00e2i \u7434/ kh\u00eem \u6709/ \u016b \u6fdf\u6fdf/ ch\u0113-ch\u0113 \u6b3e/ kh\u00f3an 'from' 'ago' 'since' 'instrument' 'has' 'many' 'appearance' Corresponding Mandarin Word(s) \u5f9e 11 w \u5c0d 12 w \u5c0d\u5b50 13 w \u5c0d\u65bc 14 w \u4ee5\u524d 21 w \u53e4\u4ee3 22 w \u53e4\u6642\u5019 23 w \u5f9e\u524d 24 w \u4ee5\uf92d 31 w \u7434 41 w \u6709 51 w \u6fdf\u6fdf 61 w \u5f88\u591a 62 w \u6a23\u5b50 71 w \u6a23\u5f0f 72 w \u6574\uf9e4 73 w 11 1 w w = 21 2 w w = 31 3 w w = 41 4 w w = 51 5 w w = 62 6 w w = 71 7 w w =", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Taiwanese", "sec_num": null }, { "text": "Assume that a particular sentence contains m words. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "words of m mn m m w w w ,..., , 2 1 . m w w w S 2 1 =", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": ", which is the most probable word sequence, is selected from the candidate words, such that P(", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "m w w w S 2 1 = ) is maximized.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "The HMM assumes that the word i w is only influenced by the previous word 1", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "\u2212 i w , thus: \u2245 = ) ( 2 1 m w w w S P \u220f = \u2212 \u00d7 m i i i w w P w P 2 1 1 ) | ( ) ( (1)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "Therefore, it searches for the word sequence", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "EQUATION", "cite_spans": [], "ref_spans": [], "eq_spans": [ { "start": 0, "end": 8, "text": "EQUATION", "ref_id": "EQREF", "raw_str": "m w w w S 2 1 = , which maximizes \u2211 = \u2212 + m i i i w w P w P 2 1 1 ) | ( log ) ( log", "eq_num": "(2)" } ], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "We use the Laplace smoothing method to solve the problem of", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "1 ( | ) 0 i i P w w \u2212 = , where no bigram of i i w w 1 \u2212", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "could be found in the training data in other words. It should be noted that the word string \u015c may not be a legal Mandarin sentence.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "In practice, we use the Viterbi algorithm to eliminate repeated computation and reduce the time complexity from exponential time to polynomial time. If a sentence S has m words, and every word has n candidate words, the time complexity will be ( ) ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Figure 2. An Example of Selecting the Best Mandarin Translation", "sec_num": null }, { "text": "We applied the Maximal Entropy Markov Model (MEMM) to the POS tag selection.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "Manning and Sch\u00fctze (1999) stated that \"Maximum entropy modeling is a framework for integrating information from many heterogeneous information sources for classification. The data for a classification problem is described as a number of features. Each feature corresponds to a constraint on the model. \u2026Choosing the maximum entropy model is motivated by the desire to preserve as much uncertainty as possible.\" MEMM includes a set of possible word and tag contexts, or \"histories\" (H), and the POS tagging set (T): are then chosen to maximize the likelihood of the training data using p:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "\u220f = = k j t h f j j t h p 1 ) , ( ) , ( \u03b1 \u03c0\u03bc (3) where , h H t T \u2208 \u2208 , \u03c0 is a normalization constant, { } k \u03b1 \u03b1 \u03bc ,...,", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "EQUATION", "cite_spans": [], "ref_spans": [], "eq_spans": [ { "start": 0, "end": 8, "text": "EQUATION", "ref_id": "EQREF", "raw_str": "\u220f \u220f \u220f = = = = = n i k j t h f j n i i i i i j t h p p L 1 1 ) , ( 1 ) , ( ) ( \u03b1 \u03c0\u03bc", "eq_num": "(4)" } ], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "As for the POS tag i t of the target word i w , we selected ten features including:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "(a) Words -five types of feature patterns: ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "2 1 1 1 2 1 , , , , + + + \u2212 \u2212 \u2212 i i i i i i i w w w w w w w . (b) POS -two types of feature patterns: 1 2 1 , \u2212 \u2212 \u2212 i i i t t t", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "1 1 2 1 , , , \u2212 \u2212 \u2212 \u2212 \u2212 \u2212 i i i i i i t t t w w w", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": ", etc. are also null (Berger et al., 1996; McCallum et al., 2000; Rabiner, 1989; Ratnaparkhi, 1996; Samuelsson, 2003; Tai, 2007; Tsai & Chen, 2004) .", "cite_spans": [ { "start": 21, "end": 42, "text": "(Berger et al., 1996;", "ref_id": "BIBREF0" }, { "start": 43, "end": 65, "text": "McCallum et al., 2000;", "ref_id": "BIBREF22" }, { "start": 66, "end": 80, "text": "Rabiner, 1989;", "ref_id": "BIBREF23" }, { "start": 81, "end": 99, "text": "Ratnaparkhi, 1996;", "ref_id": "BIBREF24" }, { "start": 100, "end": 117, "text": "Samuelsson, 2003;", "ref_id": "BIBREF25" }, { "start": 118, "end": 128, "text": "Tai, 2007;", "ref_id": "BIBREF27" }, { "start": 129, "end": 147, "text": "Tsai & Chen, 2004)", "ref_id": "BIBREF28" } ], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "In MEMM, the dependencies of observations are flexibly modeled whereas HMM assumes that observations are independent. We think MEMM is more suitable for the POS tagging task.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "We used the \"Maximum Entropy Modeling Toolkit for Python and C++\" package provided by Zhang Le to implement our system (Le, 2003) . The ten-million word POS tagged balanced Sinica corpus of the CKIP Group was used as the training data. Several million features were expanded from the ten features mentioned above, and the training time was about two days on Windows Server 2003 x64 SP2 with an Intel Xeon 3.2GHz processor (Quad-core), 8G DRAM.", "cite_spans": [ { "start": 119, "end": 129, "text": "(Le, 2003)", "ref_id": "BIBREF20" } ], "ref_spans": [], "eq_spans": [], "section": "Selecting the Most Appropriate POS According to the Corresponding Mandarin Word", "sec_num": "2.5" }, { "text": "We used the aforementioned method to perform the Taiwanese POS tagging task; nevertheless, as no standard answers were available to gauge the accuracy rate, we extracted partial results and checked them manually. The primary consideration of the manual checking procedure was the Chinese Word Segmentation and Tagging System of the CKIP group of Academia Sinica (CKIP, 2004) . We selected fourteen literary works belonging to three different eras -the Ching Dynasty, the Japanese-ruled Period, and the Post-war Era. These literary works were in the form of prose (seven), novels (five), and dramas (two). We selected the first paragraph from each composition, or, if the length (number of syllables) of the first paragraph was less than 60, we selected the second paragraph.", "cite_spans": [ { "start": 362, "end": 374, "text": "(CKIP, 2004)", "ref_id": "BIBREF4" } ], "ref_spans": [], "eq_spans": [], "section": "Results", "sec_num": "3." }, { "text": "number of tagging errors accuracy rate ( ) 100% number of total words = \u2212", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "1", "sec_num": null }, { "text": "The test data list is shown in the Appendix. Table 2 shows the test data selected for manual checking. The number of syllables, words, and incorrectly selected Mandarin words,", "cite_spans": [], "ref_spans": [ { "start": 45, "end": 52, "text": "Table 2", "ref_id": "TABREF2" } ], "eq_spans": [], "section": "1", "sec_num": null }, { "text": "as well as the POS tagging inaccuracy of each paragraph are noted.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Using Statistical Methods and Mandarin Training Data", "sec_num": null }, { "text": "A total of 1,038 words (1,496 syllables) were selected, and manual checking showed that 90 words had been incorrectly selected and 87 words were found to have inaccurate POS tagging, thus placing the average POS tagging accuracy rate at 91.6%. It should be noted that sometimes, even when the corresponding Mandarin word selected was inappropriate, the POS tagging result was still accurate. On the other hand, an appropriate or correct corresponding Mandarin word did not always have accurate POS tagging.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Using Statistical Methods and Mandarin Training Data", "sec_num": null }, { "text": "Furthermore, sometimes one Taiwanese word would correspond to two Mandarin words. For instance, while the Taiwanese word \"\u58c1\u9802/piah-t\u00e9ng\" 'on the wall' is treated as only one word, the Mandarin translation \"\u7246\u58c1 \u4e0a\" should be treated as two words. There are also occasions wherein two Taiwanese words would correspond to only one Mandarin word counterpart. For instance, the Mandarin counterpart of the Taiwanese words \"Tiong-kok/\u4e2d \u570b\" 'Chinese' and \"j\u012b/\u5b57\" 'character' was \"\u4e2d\u570b\u5b57.\" The former is processed as an unknown word, whereas the latter, which was separated into two independent words, was processed as two words. In these types of cases, if the POS tagging was accurate, we still regarded the results as accurate. If they were to be regarded as incorrect, the average accuracy rate would drop by around 2%. In Table 3 (following), examples of actual POS tagging results are shown. It is a part of id 11. In this table, the first field is the HR mixed script and POJ script (contained in brackets), and the second field is the Mandarin candidate word(s). The \"@\" symbol preceding the word indicates that no entry has been found for the Taiwanese word shown in the first field; hence the HR mixed script automatically served as the Mandarin candidate word. The third field contains the selected Mandarin word, and the final field contains the selected POS. All of the incorrectly selected Mandarin words or incorrectly selected POS tags are underlined and segregated by two asterisks \"**\" preceding the word. The correct POS tag, contained in parentheses and shown in bold type, is then added after the incorrect POS tag. ", "cite_spans": [], "ref_spans": [ { "start": 811, "end": 818, "text": "Table 3", "ref_id": "TABREF3" } ], "eq_spans": [], "section": "Using Statistical Methods and Mandarin Training Data", "sec_num": null }, { "text": "This section discusses how a more thorough check was performed to analyze the error conditions.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Error Analysis", "sec_num": "4." }, { "text": "An analysis of the errors made in the selection of Mandarin words or POS tags revealed that the selection of inappropriate Mandarin words led to POS tagging errors in 25 cases. Table 4 shows the incorrect Mandarin words selected and their respective POS. ", "cite_spans": [], "ref_spans": [ { "start": 177, "end": 184, "text": "Table 4", "ref_id": "TABREF4" } ], "eq_spans": [], "section": "Selection of Inappropriate Mandarin Word", "sec_num": "4.1" }, { "text": "There were fourteen errors made in inappropriate Mandarin word selection due to the absence of an appropriate Mandarin word in the OTMD. This also led to errors in the POS tagging. The discovery indicates the necessity of expanding the entries of the OTMD. Table 5 tabulates these errors. leh/leh \u54a7(T) '(modal particle)'", "cite_spans": [], "ref_spans": [ { "start": 257, "end": 264, "text": "Table 5", "ref_id": "TABREF5" } ], "eq_spans": [], "section": "Absence of Appropriate Mandarin Translation in OTMD", "sec_num": "4.2" }, { "text": "\u5728(P) 'doing' 3 times \u715e/soah \u7d50\u675f(VHC) 'finish' \u537b(D) 'but' teh/teh \u5728(P) '(an indicator or location)' \u8457(Di) '(an adverbial particle)' \u9802/t\u00e9ng \u9802(VC) 'lift' \u4e0a(Nes) '(the first half part)' ti\u0101\u207f-ti\u0101\u207f / ti\u0101\u207f-ti\u0101\u207f \u5e38\u5e38(D) 'often' \u800c\u5df2(T) 'just'", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Absence of Appropriate Mandarin Translation in OTMD", "sec_num": "4.2" }, { "text": "\u8f49 / t\u0144g \u8abf\u89e3(VC) 'mediate' \u8f49(VAC) 'turn'", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Absence of Appropriate Mandarin Translation in OTMD", "sec_num": "4.2" }, { "text": "Ten of the POS tagging errors were made because the word was an unknown word. Parts of these unknown words correspond to two Mandarin words. These unknown words are tabulated in Table 6 . ", "cite_spans": [], "ref_spans": [ { "start": 178, "end": 185, "text": "Table 6", "ref_id": "TABREF6" } ], "eq_spans": [], "section": "Unknown Words from the Viewpoint of Mandarin", "sec_num": "4.3" }, { "text": "Five of the POS tagging errors were probably due to the occurrence of a previous POS tagging error. These are categorized as propagation errors and include one unknown word.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Propagation Error", "sec_num": "4.4" }, { "text": "The personal name \"\u5929\u8cdc\" of \"\u5929\u8cdc ah/Thian-s\u00f9 ah\" (not an unknown word) which has been tagged as \"A\" with the suffix \"ah\" tagged as \"T\" or \"Di\" (which appeared twice in all; once, the selected Mandarin word was \"\u554a\" and in other instance it was \"\uf9ba\").", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Other Cases", "sec_num": "4.5" }, { "text": "The Taiwanese word \"\u5c0d/t\u00f9i\" under general circumstances is synonymous with the Mandarin word \"\u5f9e\"'from.' This word appeared ten times in the test data. The system selected the Mandarin word \"\u5c0d\"'for' eight times and the word \"\u5f9e\" twice for its counterpart. Nevertheless, under both circumstances, the POS tag of the word was always \"P\"; thus, the different word choice did not affect the accuracy of the POS tagging.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Other Cases", "sec_num": "4.5" }, { "text": "There were also 30 errors made that leave us unable to clearly explain the reasons. Table 7 lists some examples. ", "cite_spans": [], "ref_spans": [ { "start": 84, "end": 92, "text": "Table 7", "ref_id": "TABREF7" } ], "eq_spans": [], "section": "Other Cases", "sec_num": "4.5" }, { "text": "A summary of the causes of the errors made during the POS tagging and their frequency percentages is tabulated in Table 8 . ", "cite_spans": [], "ref_spans": [ { "start": 114, "end": 121, "text": "Table 8", "ref_id": "TABREF8" } ], "eq_spans": [], "section": "Summary of Error Conditions", "sec_num": "4.6" }, { "text": "The ideal situation would be to resolve the foregoing errors and use this method to conduct the Taiwanese POS tagging to achieve an accuracy rate of 97.1%. Nevertheless, there is an apparent difficulty in the realization of this goal.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Is Improvement Possible?", "sec_num": "5.1" }, { "text": "There are differences between the Taiwanese word order and the Mandarin word order; thus, the selection of the incorrect Mandarin word, and consequently incorrect POS tagging, occurred with high probability. The absence of appropriate Mandarin translation was the second leading cause of the POS tagging errors.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Is Improvement Possible?", "sec_num": "5.1" }, { "text": "The unknown word problem was also a cause of POS tagging errors. From the Mandarin perspective, these words are not actually unknown words; this problem mostly resulted from the fact that translations between different languages are not one-to-one mappings. Another significant factor involves the use of hyphens in the POJ script, as their usage has not yet been standardized. It is probable that due to the use of Han characters, word boundaries are relatively vague in the different languages of the Chinese language family.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Is Improvement Possible?", "sec_num": "5.1" }, { "text": "In Taiwanese, some words take on the POJ script, thus, the use of the hyphen. Used one way, they separate the syllables of words, making it possible for a syllable to correspond to a Han character; used another way, they serve as word separators. Each syllable in a hyphenated word represents a unigram, and a space separates each word. Unfortunately, no original word boundaries of Han character writing can be found to correspond to the hyphenated word.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Hyphen Problems, Distinction between Taiwanese and Mandarin", "sec_num": "5.2" }, { "text": "In addition, Taiwanese has around 3,000 legal syllables, whereas Mandarin has around 1,200 legal syllables (Chan, 2008) . Because of this, it may be said that the Taiwanese language has more single-syllable words. Nevertheless, as a single-syllable word may have several corresponding Han characters, the use of two-syllable or multi-syllable words resolves most of the problems.", "cite_spans": [ { "start": 107, "end": 119, "text": "(Chan, 2008)", "ref_id": "BIBREF1" } ], "ref_spans": [], "eq_spans": [], "section": "Hyphen Problems, Distinction between Taiwanese and Mandarin", "sec_num": "5.2" }, { "text": "For instance, if the Taiwanese word \"\u9019\u500b\"'this one' is written as \"chit \u00ea\" (no hyphen used), the syllable \"chit\" may be made to correspond to several Mandarin words, such as \"\u9019\" 'this,' \"\u8077\"'job,' \"\u8cea\" 'quality,' \"\u7e54,\" 'knit,' etc. The syllable \"\u00ea\" may also be made to correspond to several Mandarin words, such as \"\u7684\" 'of,' \"\u500b\" '(a numerary adjunct),' \"\u978b\" 'shoe,' etc. If the word is written as \"chit-\u00ea\" (hyphenated), it definitely corresponds to \"\u9019\u500b\" in HR script. Hence, under the POJ script, the writer may tend to use a hyphen to link a single-syllable word to another single-syllable word if these two single-syllable words may likely form one composite word or one phrase. Present practices show that the word \"\u9019\u500b\" may appear hyphenated or in a separated syllable form, thus creating inconsistencies.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Hyphen Problems, Distinction between Taiwanese and Mandarin", "sec_num": "5.2" }, { "text": "As the use of hyphenated words creates the problem of one Taiwanese word corresponding to two Mandarin words, if the original text is not revised and the Mandarin corresponding word is manifested as an unknown word, it may be possible to just remove the hyphen and try again. This method may reduce the chance of POS tagging errors due to the unknown word factor.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Hyphen Problems, Distinction between Taiwanese and Mandarin", "sec_num": "5.2" }, { "text": "We investigated whether texts of a different era or a different literary genre would affect the accuracy rate of the POS tagging. Table 10 shows the POS tagging accuracy rates for texts of three types of literary genres and Table 11 shows the POS tagging accuracy rates for texts of literary works belonging to three different periods or eras. Table 9 shows that the POS tagging accuracy rate for novel materials is comparably lower than other genres; whereas Table 10 indicates that the POS tagging accuracy rate for the materials written in the Post-war era are comparably lower than the other periods investigated. Basically, there are no significant differences among three genres or three eras as a whole. After deliberation, we found that the individual writing style of authors is actually the dominant factor of the POS tagging accuracy. From Table 2 , the individual POS tagging accuracy varies from 83.6% to 98.0%.", "cite_spans": [], "ref_spans": [ { "start": 130, "end": 138, "text": "Table 10", "ref_id": "TABREF0" }, { "start": 224, "end": 232, "text": "Table 11", "ref_id": "TABREF0" }, { "start": 344, "end": 351, "text": "Table 9", "ref_id": null }, { "start": 460, "end": 468, "text": "Table 10", "ref_id": "TABREF0" }, { "start": 851, "end": 858, "text": "Table 2", "ref_id": "TABREF2" } ], "eq_spans": [], "section": "The Distinction between Different Eras or Different Genres", "sec_num": "5.3" }, { "text": "We proposed a Taiwanese POS tagging method using a statistical method and Mandarin training data, and we achieved an accuracy rate of 91.6%. Due to the lack of Taiwanese training data, we sought the help of Mandarin.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Conclusion and Future Works", "sec_num": "6." }, { "text": "This strategy could also be applied to other languages that lack resources. We think that this is a very important idea. It is preferable to select an intermediate language close to the target language from the viewpoint of the language family.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Conclusion and Future Works", "sec_num": "6." }, { "text": "We also developed an online Taiwanese word segmentation and POS tagging system for people who are interested in this topic. Users can input Taiwanese text and get the POS tagging results. It is somewhat difficult for a user to prepare both POJ and HR mixed scripts; therefore, we also provide the functions in the absence of one of these two scripts (Iunn, et. al., 2007) . This, however, will decrease the accuracy rate.", "cite_spans": [ { "start": 350, "end": 371, "text": "(Iunn, et. al., 2007)", "ref_id": "BIBREF14" } ], "ref_spans": [], "eq_spans": [], "section": "Conclusion and Future Works", "sec_num": "6." }, { "text": "If we can construct a Taiwanese-Mandarin parallel corpus, we can use other methods like the Coerced Markov Models proposed by Fung and Wu (1995) to accomplish the Taiwanese POS tagging task.", "cite_spans": [ { "start": 126, "end": 144, "text": "Fung and Wu (1995)", "ref_id": "BIBREF6" } ], "ref_spans": [], "eq_spans": [], "section": "Conclusion and Future Works", "sec_num": "6." }, { "text": "We hope that we can proceed to the construction of Taiwanese Treebank.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Conclusion and Future Works", "sec_num": "6." } ], "back_matter": [ { "text": "This research was supported in part by National Science Council of Taiwan, under the contract number NSC 95-2221-E-122 -006. Thanks to the National Museum of Taiwanese Literature of Taiwan for providing plentiful written Taiwanese literature data. We also thank the anonymous reviewers for their constructive opinions.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Acknowledgments", "sec_num": null } ], "bib_entries": { "BIBREF0": { "ref_id": "b0", "title": "A Maximum Entropy Approach to Natural Language Processing", "authors": [ { "first": "A", "middle": [ "L" ], "last": "Berger", "suffix": "" }, { "first": "S", "middle": [ "A D" ], "last": "Pietra", "suffix": "" }, { "first": "V", "middle": [ "J D" ], "last": "Pietra", "suffix": "" } ], "year": 1996, "venue": "Computational Linguistics", "volume": "22", "issue": "1", "pages": "39--71", "other_ids": {}, "num": null, "urls": [], "raw_text": "Berger, A. L., Pietra, S. A. D., & Pietra, V. J. D. (1996). A Maximum Entropy Approach to Natural Language Processing. Computational Linguistics, 22(1), 39-71.", "links": null }, "BIBREF1": { "ref_id": "b1", "title": "Comparison with the Usage of Academic and Non-academic Taiwanese Words", "authors": [ { "first": "K", "middle": [ "I" ], "last": "Chan", "suffix": "" } ], "year": 2008, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Chan, K. I. (2008). Comparison with the Usage of Academic and Non-academic Taiwanese Words. Master thesis, National Taitung University.", "links": null }, "BIBREF2": { "ref_id": "b2", "title": "T3 Taiwanese Treebank and Brill Part-of-Speech Tagger. Master thesis", "authors": [ { "first": "S", "middle": [ "Y" ], "last": "Chou", "suffix": "" } ], "year": 2006, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Chou, S. Y. (2006). T3 Taiwanese Treebank and Brill Part-of-Speech Tagger. Master thesis, Hsin-chu: National Tsing Hua University.", "links": null }, "BIBREF3": { "ref_id": "b3", "title": "Analysis of Chinese Part-of-speech", "authors": [ { "first": "", "middle": [], "last": "Ckip", "suffix": "" } ], "year": 1993, "venue": "The Association for Computational Linguistics and Chinese Language", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "CKIP. (1993). Analysis of Chinese Part-of-speech. The Association for Computational Linguistics and Chinese Language.", "links": null }, "BIBREF4": { "ref_id": "b4", "title": "Chinese Word Segmentation and Tagging System", "authors": [ { "first": "", "middle": [], "last": "Ckip", "suffix": "" } ], "year": 2004, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "CKIP. (2004). Chinese Word Segmentation and Tagging System, (Retrieved 2009/4/10) http://ckipsvr.iis.sinica.edu.tw/.", "links": null }, "BIBREF5": { "ref_id": "b5", "title": "A dictionary of Southern Min '\u53f0\u82f1\u8fad\u5178", "authors": [ { "first": "B", "middle": [ "L M" ], "last": "Embree", "suffix": "" } ], "year": 1984, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Embree, B. L. M. (1984). A dictionary of Southern Min '\u53f0\u82f1\u8fad\u5178,' Taipei: Taipei Language Institute.", "links": null }, "BIBREF6": { "ref_id": "b6", "title": "Coerced Markov Models for cross-lingual lexical tag relations", "authors": [ { "first": "P", "middle": [], "last": "Fung", "suffix": "" }, { "first": "D", "middle": [ "K" ], "last": "Wu", "suffix": "" } ], "year": 1995, "venue": "the 6 th International Conference on Theoretical and Methodological Issues in Machine Translation", "volume": "1", "issue": "", "pages": "240--255", "other_ids": {}, "num": null, "urls": [], "raw_text": "Fung, P., & Wu, D. K. (1995). Coerced Markov Models for cross-lingual lexical tag relations. in the 6 th International Conference on Theoretical and Methodological Issues in Machine Translation, 1, 1995, Leuven, Belgium, 240-255.", "links": null }, "BIBREF7": { "ref_id": "b7", "title": "Ethnologue: Languages of the world", "authors": [ { "first": "R", "middle": [ "G" ], "last": "Gordon", "suffix": "" } ], "year": 2005, "venue": "Dallas:SIL International", "volume": "15", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Gordon, R. G. Jr. ed. (2005). Ethnologue: Languages of the world (15 th ed.). Dallas:SIL International.", "links": null }, "BIBREF8": { "ref_id": "b8", "title": "Language, Society and Ethnicity", "authors": [ { "first": "S", "middle": [ "F" ], "last": "Huang", "suffix": "" } ], "year": 1995, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Huang, S. F.(1995). Language, Society and Ethnicity (2 nd ed.). Taipei: Crane.", "links": null }, "BIBREF9": { "ref_id": "b9", "title": "Online Taiwanese-Mandarin Dictionary", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" } ], "year": 2000, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G. (2000). Online Taiwanese-Mandarin Dictionary. (Retrieved 2009/4/10) http://iug.csie.dahan.edu.tw/q/q.asp.", "links": null }, "BIBREF10": { "ref_id": "b10", "title": "Online Taiwanese Syllable Dictionary", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" } ], "year": 2003, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G. (2003a). Online Taiwanese Syllable Dictionary. (Retrieved 2009/4/10) http://iug.csie.dahan.edu.tw/TG/jitian/.", "links": null }, "BIBREF11": { "ref_id": "b11", "title": "Survey of the Online Taiwanese-Mandarin Dictionary--Discussion of Building Technique and its Utilization", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" } ], "year": 2003, "venue": "the Proceedings of 3 rd International Conference on Internet Chinese Education", "volume": "", "issue": "", "pages": "132--141", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G. (2003b). Survey of the Online Taiwanese-Mandarin Dictionary--Discussion of Building Technique and its Utilization. in the Proceedings of 3 rd International Conference on Internet Chinese Education, 2003b, Overseas Chinese Affairs Commission, 132-141.", "links": null }, "BIBREF12": { "ref_id": "b12", "title": "Online Taiwanese Concordancer System", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" } ], "year": 2003, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G. (2003c). Online Taiwanese Concordancer System. (Retrieved 2009/4/10) http://iug.csie.dahan.edu.tw/TG/concordance/.", "links": null }, "BIBREF13": { "ref_id": "b13", "title": "Taiwanese Corpus Collection and Corpus Based Syllable / Word Frequency Counts for Written Taiwanese", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" } ], "year": 2005, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G. (2005). Taiwanese Corpus Collection and Corpus Based Syllable / Word Frequency Counts for Written Taiwanese. (Retrieved 2009/4/10)", "links": null }, "BIBREF14": { "ref_id": "b14", "title": "New Manifestation of the Taiwanese vernacular literature --Introduction to Digital Archive for Written Taiwanese. National Museum of Taiwanese Literature Communication", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" } ], "year": 2007, "venue": "", "volume": "15", "issue": "", "pages": "42--44", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G. (2007). New Manifestation of the Taiwanese vernacular literature --Introduction to Digital Archive for Written Taiwanese. National Museum of Taiwanese Literature Communication, 15, 42-44.", "links": null }, "BIBREF15": { "ref_id": "b15", "title": "Processing Techniques for Written Taiwanese--Tone Sandhi and POS Tagging", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" } ], "year": 2009, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G. (2009). Processing Techniques for Written Taiwanese--Tone Sandhi and POS Tagging. PhD thesis, National Taiwan University.", "links": null }, "BIBREF16": { "ref_id": "b16", "title": "Introduction to online Taiwanese Dictionaries and Corpora. Language, Society and Culture Series 2: Multiculturalism Thinking of the Language Policy", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" }, { "first": "K", "middle": [ "G" ], "last": "Lau", "suffix": "" } ], "year": 2007, "venue": "", "volume": "", "issue": "", "pages": "311--328", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G., & Lau, K. G. (2007). Introduction to online Taiwanese Dictionaries and Corpora. Language, Society and Culture Series 2: Multiculturalism Thinking of the Language Policy, 2007, Institute of Linguistics of Academia Sinica, 311-328.", "links": null }, "BIBREF17": { "ref_id": "b17", "title": "Modeling Taiwanese Southern-Min Tone Sandhi Using Rule-Based Methods", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" }, { "first": "K", "middle": [ "G" ], "last": "Lau", "suffix": "" }, { "first": "H", "middle": [ "G" ], "last": "Tan-Tenn", "suffix": "" }, { "first": "S", "middle": [ "A" ], "last": "Lee", "suffix": "" }, { "first": "C", "middle": [ "Y" ], "last": "Kao", "suffix": "" } ], "year": 2007, "venue": "International Using Statistical Methods and Mandarin Training Data Journal of Computational Linguistics and Chinese Language Processing", "volume": "12", "issue": "4", "pages": "349--370", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G., Lau, K. G., Tan-Tenn, H. G., Lee, S. A., & Kao, C. Y. (2007). Modeling Taiwanese Southern-Min Tone Sandhi Using Rule-Based Methods. International Using Statistical Methods and Mandarin Training Data Journal of Computational Linguistics and Chinese Language Processing, 12(4), 349-370.", "links": null }, "BIBREF18": { "ref_id": "b18", "title": "Online Taiwanese Word Segmentation and POS Tagging System", "authors": [ { "first": "U", "middle": [ "G" ], "last": "Iunn", "suffix": "" }, { "first": "K", "middle": [ "G" ], "last": "Lau", "suffix": "" }, { "first": "C", "middle": [ "H" ], "last": "Tai", "suffix": "" } ], "year": 2007, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Iunn, U. G., Lau, K. G., & Tai, C. H. (2007). Online Taiwanese Word Segmentation and POS Tagging System, (Retrieved 2009/4/10) http://iug.csie.dahan.edu.tw/TGB/tagging/tagging.asp.", "links": null }, "BIBREF19": { "ref_id": "b19", "title": "Finding Mandarin Candidate Words by POJ script and Han-Romanization mixed script word pair", "authors": [ { "first": "K", "middle": [ "G" ], "last": "Lau", "suffix": "" } ], "year": 2007, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Lau, K. G. (2007). Finding Mandarin Candidate Words by POJ script and Han-Romanization mixed script word pair, (Retrieved 2009/4/10)", "links": null }, "BIBREF20": { "ref_id": "b20", "title": "Maximum Entropy Modeling Toolkit for Python and C++", "authors": [ { "first": "Z", "middle": [], "last": "Le", "suffix": "" } ], "year": 2003, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Le, Z. (2003). Maximum Entropy Modeling Toolkit for Python and C++. (Retrieved 2009/7/10) http://homepages.inf.ed.ac.uk/lzhang10/maxent_toolkit.html.", "links": null }, "BIBREF21": { "ref_id": "b21", "title": "Foundations of Statistical Natural Language Processing", "authors": [ { "first": "C", "middle": [], "last": "Manning", "suffix": "" }, { "first": "H", "middle": [], "last": "Sch\u00fctze", "suffix": "" } ], "year": 1999, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Manning, C., & Sch\u00fctze, H. (1999). Foundations of Statistical Natural Language Processing, Cambridge: MIT Press.", "links": null }, "BIBREF22": { "ref_id": "b22", "title": "Maximum Entropy Markov Models for Information Extraction and Segmentation", "authors": [ { "first": "A", "middle": [], "last": "Mccallum", "suffix": "" }, { "first": "D", "middle": [], "last": "Freitag", "suffix": "" }, { "first": "F", "middle": [], "last": "Pereira", "suffix": "" } ], "year": 2000, "venue": "the Proceedings of 17 th International Conference on Machine Learning", "volume": "", "issue": "", "pages": "591--598", "other_ids": {}, "num": null, "urls": [], "raw_text": "McCallum, A., Freitag, D., & Pereira, F. (2000). Maximum Entropy Markov Models for Information Extraction and Segmentation. in the Proceedings of 17 th International Conference on Machine Learning, Stanford University, 591-598.", "links": null }, "BIBREF23": { "ref_id": "b23", "title": "A tutorial on hidden Markov models and selected applications in speech recognition", "authors": [ { "first": "L", "middle": [ "R" ], "last": "Rabiner", "suffix": "" } ], "year": 1989, "venue": "the Proceedings of the IEEE", "volume": "77", "issue": "", "pages": "257--286", "other_ids": {}, "num": null, "urls": [], "raw_text": "Rabiner, L. R. (1989). A tutorial on hidden Markov models and selected applications in speech recognition. in the Proceedings of the IEEE, 77, 257-286.", "links": null }, "BIBREF24": { "ref_id": "b24", "title": "A maximum entropy model for part-of-speech tagging", "authors": [ { "first": "A", "middle": [], "last": "Ratnaparkhi", "suffix": "" } ], "year": 1996, "venue": "Proceedings of Conference on Empirical Methods in Natural Language Processing", "volume": "", "issue": "", "pages": "133--142", "other_ids": {}, "num": null, "urls": [], "raw_text": "Ratnaparkhi, A. (1996). A maximum entropy model for part-of-speech tagging. Proceedings of Conference on Empirical Methods in Natural Language Processing, University of Pennsylvania, 133-142.", "links": null }, "BIBREF25": { "ref_id": "b25", "title": "Statistical methods. in the Oxford Handbook of Computational Linguistics", "authors": [ { "first": "C", "middle": [], "last": "Samuelsson", "suffix": "" } ], "year": 2003, "venue": "", "volume": "", "issue": "", "pages": "358--375", "other_ids": {}, "num": null, "urls": [], "raw_text": "Samuelsson, C. (2003). Statistical methods. in the Oxford Handbook of Computational Linguistics, Oxford University Press, 358-375.", "links": null }, "BIBREF26": { "ref_id": "b26", "title": "T3 Taiwanese Treebank and Brill Parser", "authors": [ { "first": "D", "middle": [ "M" ], "last": "Shi", "suffix": "" } ], "year": 2006, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Shi, D. M. (2006). T3 Taiwanese Treebank and Brill Parser, Master thesis, Hsin-chu: National Tsing Hua University.", "links": null }, "BIBREF27": { "ref_id": "b27", "title": "Word and POS tagging selection for Taiwanese Language", "authors": [ { "first": "C", "middle": [ "H" ], "last": "Tai", "suffix": "" } ], "year": 2007, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Tai, C. H. (2007). Word and POS tagging selection for Taiwanese Language, (Retrieved 2009/4/10) http://140.109.19.105/.", "links": null }, "BIBREF28": { "ref_id": "b28", "title": "1 1885 prose Reverend Ia\u030d p '\uf96e \u7267\u5e2b' Pe\u030d h-\u014de-j\u012b \u00ea l\u012b-ek 'The Benefits of Using Pe\u030d h-\u014de-j\u012b, \u767d\u8a71\u5b57\u7684\uf9dd\u76ca' 162 2 1893 prose Reverend Kam '\u7518 \u7267\u5e2b' Chhi\u207f-m\u00ee o\u030d h 'Blind Study", "authors": [ { "first": "Y", "middle": [ "F" ], "last": "Tsai", "suffix": "" }, { "first": "K", "middle": [ "J" ], "last": "Chen", "suffix": "" } ], "year": 2004, "venue": "International Journal of Computational Linguistics and Chinese Language Processing", "volume": "9", "issue": "1", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Tsai, Y. F., & Chen, K. J. (2004). Reliable and Cost-Effective Pos-Tagging. International Journal of Computational Linguistics and Chinese Language Processing, 9(1), 2004, 83-96. 1 1885 prose Reverend Ia\u030d p '\uf96e \u7267\u5e2b' Pe\u030d h-\u014de-j\u012b \u00ea l\u012b-ek 'The Benefits of Using Pe\u030d h-\u014de-j\u012b, \u767d\u8a71\u5b57\u7684\uf9dd\u76ca' 162 2 1893 prose Reverend Kam '\u7518 \u7267\u5e2b' Chhi\u207f-m\u00ee o\u030d h 'Blind Study, \u9752\u7791\u5b78' 66", "links": null }, "BIBREF29": { "ref_id": "b29", "title": "\u9707\u707d\u8a18' 122 5 1954 prose \u00d4 . B\u00fbn-t\u00ee '\u80e1\u6587\u6c60' T\u014da-soa\u207f chhi\u00f9\u207f-koa 'A High Mountains sing, \u5927\u5c71\u5531\u6b4c' 74 6 1990 prose T\u00e2n G\u012b-j\u00een '\u9673\u7fa9 \u4ec1' L\u0101u-l\u00e2ng \u00ea k\u00e8-ta\u030d t 'The Value of The Elderly People, \uf934\u4eba\u7684\u50f9\u503c' 75 7 2000 prose T\u00e2n B\u00eang-j\u00een '\u9673\u660e \u4ec1' S\u00fbn-ch\u00eang \u00d4ng P\u00f3-chhoan 'Pure Love \u00d4ng P\u00f3-chhoan, \u7d14\u60c5\u738b\u5bf6\u91e7' 112 8 1890 novel Unknown An-lo\u030d k-ke 'Safety and Happiness Street, \u5b89\uf914\u8857' 101 9 1924 novel L\u014da J\u00een-seng '\u8cf4\u4ec1 \u8072' \u00c1n-ni\u00e1 \u00ea Ba\u030d k-s\u00e1i 'Mother's Tears, \u6bcd\u89aa \u7684\u773c\uf94d' 133 10 1955 novel N\u011d H\u00f4ai-un '\u9ec3\u61f7 \u6069' Chh\u00e1u-tui t\u00e9ng \u00ea b\u00een-b\u0101ng 'Dreams on the Grass Stack", "authors": [ { "first": "", "middle": [], "last": "Prose H S K Ph\u00edn-H\u0113ng \u00ca \u00dbi-Th\u00f4an", "suffix": "" } ], "year": null, "venue": "Inheritance of Morality, \u54c1\ufa08\u7684\u907a\u50b3' 180 4 1935 prose Ong Chong-t\u00eang '\u6c6a\u5b97\u7a0b' Ch\u00edn-chai k\u00ec 'Earthquake Disaster Record", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "prose H S K Ph\u00edn-h\u0113ng \u00ea \u00fbi-th\u00f4an 'Inheritance of Morality, \u54c1\ufa08\u7684\u907a\u50b3' 180 4 1935 prose Ong Chong-t\u00eang '\u6c6a\u5b97\u7a0b' Ch\u00edn-chai k\u00ec 'Earthquake Disaster Record, \u9707\u707d\u8a18' 122 5 1954 prose \u00d4 . B\u00fbn-t\u00ee '\u80e1\u6587\u6c60' T\u014da-soa\u207f chhi\u00f9\u207f-koa 'A High Mountains sing, \u5927\u5c71\u5531\u6b4c' 74 6 1990 prose T\u00e2n G\u012b-j\u00een '\u9673\u7fa9 \u4ec1' L\u0101u-l\u00e2ng \u00ea k\u00e8-ta\u030d t 'The Value of The Elderly People, \uf934\u4eba\u7684\u50f9\u503c' 75 7 2000 prose T\u00e2n B\u00eang-j\u00een '\u9673\u660e \u4ec1' S\u00fbn-ch\u00eang \u00d4ng P\u00f3-chhoan 'Pure Love \u00d4ng P\u00f3-chhoan, \u7d14\u60c5\u738b\u5bf6\u91e7' 112 8 1890 novel Unknown An-lo\u030d k-ke 'Safety and Happiness Street, \u5b89\uf914\u8857' 101 9 1924 novel L\u014da J\u00een-seng '\u8cf4\u4ec1 \u8072' \u00c1n-ni\u00e1 \u00ea Ba\u030d k-s\u00e1i 'Mother's Tears, \u6bcd\u89aa \u7684\u773c\uf94d' 133 10 1955 novel N\u011d H\u00f4ai-un '\u9ec3\u61f7 \u6069' Chh\u00e1u-tui t\u00e9ng \u00ea b\u00een-b\u0101ng 'Dreams on the Grass Stack, \u8349\u5806\u4e0a\u7684\u5922' 116 11 1990 novel I\u00fb\u207f \u00dan-gi\u00e2n '\u694a\u5141 \u8a00' translated", "links": null }, "BIBREF30": { "ref_id": "b30", "title": "H\u00e1i-ph\u012b\u207f Sin-ni\u00fb 'Bride on The Cape, \u5cac\u89d2\u4e0a\u7684\u65b0\u5a18' 94 12 2006 novel L\u00e2u S\u00eang-hi\u00e2n '\uf9c7 \u627f\u8ce2' Chia\u030d h-ch\u014de 'Plead Guilty, \u4f0f\u7f6a' 92 13 1924 drama L\u00eem B\u014d . -seng '\uf9f4 \u8302\u751f' H\u00ec-chhut: L\u014d . -tek k\u00e1i k\u00e0u ' Drama: Ruth Reformed Church, \u6232\u9f63:\uf937\u5f97\u6539\u6559' 77 14 1950 drama T\u00e2n Chheng-tiong '\u9673\u6e05\u5fe0' translated", "authors": [], "year": null, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "H\u00e1i-ph\u012b\u207f Sin-ni\u00fb 'Bride on The Cape, \u5cac\u89d2\u4e0a\u7684\u65b0\u5a18' 94 12 2006 novel L\u00e2u S\u00eang-hi\u00e2n '\uf9c7 \u627f\u8ce2' Chia\u030d h-ch\u014de 'Plead Guilty, \u4f0f\u7f6a' 92 13 1924 drama L\u00eem B\u014d . -seng '\uf9f4 \u8302\u751f' H\u00ec-chhut: L\u014d . -tek k\u00e1i k\u00e0u ' Drama: Ruth Reformed Church, \u6232\u9f63:\uf937\u5f97\u6539\u6559' 77 14 1950 drama T\u00e2n Chheng-tiong '\u9673\u6e05\u5fe0' translated", "links": null }, "BIBREF31": { "ref_id": "b31", "title": "Venice \u00ea Seng-l\u00ed-l\u00e2ng 'Venice Businessman", "authors": [], "year": null, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Venice \u00ea Seng-l\u00ed-l\u00e2ng 'Venice Businessman, \u5a01\u5c3c\u65af\u7684\u751f\u610f\u4eba' 92", "links": null }, "BIBREF32": { "ref_id": "b32", "title": "Note: the original author of id 11 is S\u00f2ng Te\u030d k-l\u00e2i '\u5b8b\u6fa4\u840a", "authors": [], "year": null, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Note: the original author of id 11 is S\u00f2ng Te\u030d k-l\u00e2i '\u5b8b\u6fa4\u840a,' id 14 is Shakespeare", "links": null } }, "ref_entries": { "FIGREF0": { "uris": null, "num": null, "text": "Figure 1 shows our system architecture diagram.", "type_str": "figure" }, "FIGREF3": { "uris": null, "num": null, "text": "are designated to manipulate the unknown words. If i w is an unknown word, we segment i w with a maximal matching strategy; thus, three morpheme features are set to null. Moreover, if i w is at the beginning or end of a sentence, certain features are likewise given a null value. For instance, when i=1, the feature values of 1 2", "type_str": "figure" }, "TABREF0": { "html": null, "content": "
HR Mixed ScriptPOJ ScriptMandarin Translation
ch\u014danch\u014dan\u626d
\u64b0ch\u014dan\u4e0a
", "num": null, "type_str": "table", "text": "" }, "TABREF2": { "html": null, "content": "
idNo. of SyllablesNo. of WordsErrorsTagging errorsAccuracy rate(%)
11621099694.5
266464393.5
31801196893.3
4122883693.2
574514198.0
675497785.7
711287131286.2
8101777988.3
9133937990.3
10116823396.3
1194597591.5
129261101083.6
1377598493.2
1492583493.1
Totally1,4961,038908791.6
", "num": null, "type_str": "table", "text": "" }, "TABREF3": { "html": null, "content": "
TaiwaneseMandarin Candidate WordsSelected WordPOS tagging
\u6211[g\u00f3a] 'I'\u6211\u6211Nh
\u5c07[chiong] 'let'\u5c07\u5c07D
\u8349\u5e3d\u4ed4[chh\u00e1u-b\u014d-\u00e1] 'straw hat' @\u8349\u5e3d\u4ed4\u8349\u5e3d\u4ed4Na
\u639b[k\u00f2a] 'hang'\u5e36;\u639b;\u6234**\u5e36 \u639bVC
t\u012b [t\u012b] 'at'\u5728\u5728P
\u58c1\u9802[piah-t\u00e9ng] 'on the wall'\u7246\u58c1\u4e0a\u7246\u58c1\u4e0aNc
\uff0c[,]\uff0c\uff0cCommaCategory
\ufa08\uf9e1[h\u00eang-l\u00ed] 'baggage'\ufa08\uf9e1\ufa08\uf9e1Na
kh\u00eang[kh\u00eang] 'arrange'\u6536\uf973;\u76e4\u9ede\u6536\uf973VC
kh\u00eang[kh\u00eang] 'arrange'\u6536\uf973;\u76e4\u9ede\u6536\uf973VC
leh[leh] '(modal particle)'\u54a7\u54a7T
\uff0c[,]\uff0c\uff0cCommaCategory
\u5750[ch\u0113] 'sit'\u5750\u5750VA
t\u00f2a[t\u00f2a] 'at'\u4f4f**\u4f4f**VCL(P)
\u5c0f\u5e97[si\u00f3-ti\u00e0m] 'store'@\u5c0f\u5e97\u5c0f\u5e97Na
\u00ea[\u00ea] 'of'\u7684\u7684DE
tha-th\u00e1-m\u00ec[tha-tha-m\u00ec] 'tatami'\u584c\u584c\u7c73\u584c\u584c\u7c73Na
\u9802 k\u00f4an[t\u00e9ng-k\u00f4an] 'above'\u4e0a\u9762\u4e0a\u9762Ncd
\uff0c[,]\uff0c\uff0cCommaCategory
\u770b[kh\u00f2a\u207f] 'see'\u770b\u770bVC
\u7a97\u5916[thang-g\u014da] 'outside the@\u7a97\u5916\u7a97\u5916Nc
window'
", "num": null, "type_str": "table", "text": "" }, "TABREF4": { "html": null, "content": "
WordSelected Mandarin word and POSMore appropriate Mandarin word and POSRemark
\u62bc/ah\u5f37\u5236(D) 'compel'\u62bc(VC) 'take into custody'
bat/bat\u77e5\u9053(VK) 'know'\u66fe(D) 'ever'
\u7121/b\u00f4\uf967(D) 'not'\u6c92\u6709(VJ) 'not have'2 times
chham/chham\u548c(P) 'and'\u647b(VC) 'accompany'
\u9032\u524d /ch\u00ecn-ch\u00eang\u4e4b\u524d(Ng) 'before'\u5411 \u524d(P Nes) 'forward'
\u9019\u865f/chit-h\u014d\u9019\u6a23(VH) 'such'\u9019\u7a2e(Nep Nf) 'this kind of'2 times
\u8f49/ch\u014dan\u4e0a(Ncd) 'above'\u8f49(Vac) 'turn'2 times
\u5916/g\u014da\u5916(Ng) 'outside'\u958b\u5916(Neqa) 'more'2 times
\u592d\u58fd/i\u00e1u-si\u016b\u975e\u5e38(Dfa) 'very'\u65e9\u592d(VH) 'dead early'
\u52a0/ke\u4e0a(Ncd) 'above'\u591a(Dfa) 'more'
\u50f9\u503c/k\u00e8-ta\u030d t\u503c\u5f97(VH) 'worthy'\u50f9\u503c(Na) 'value'
\u8173/kha\u500b(DE) '(a numerary adjunct)'\u4e0b(Ncd) 'under'
\u9ec3 h\u00f3a\u207f/n\u011d-h\u00f3a\u207f \u7f55(D) 'rarely'\u6dfa\u9ec3(A) 'light yellow'
\u501a/\u00f3a\u4f9d(P) 'in accordance with'\u9760(VJ) 'lean against'
\u6d3b/o\u030d ah\u751f\u6d3b(Na) 'life'\u6d3b(VH) 'live'
\u7834\u76f8/ph\u00f2a-si\u00f9\u207f \u7834(VHC) 'break'\u6b98\u5ee2(Na) 'disabled'
\u7d30\u6f22/s\u00e8-h\u00e0n\u5c0f\u6642\u5019(Nd) 'in one's childhood'\uf98e\u5e7c(VH) 'young'
\u76f8\u501f\u554f /sio-chioh-mn\u1e21\u62db\u547c(VC) 'greet'\u6253\u62db\u547c(VB) 'say hello'
\u642d/tah\u5730\u65b9(Na) 'location'\u642d(VC) 'construct'
tio\u030d h/tio\u030d h\u5c31(P) '(an auxiliary confirming and stressing the verb following)'\u8457(VCL) 'come into contact with'
\u8457/tio\u030d h\u5c31(P) '(an auxiliary confirming and stressing the verb following)'\u5f97(D) 'need to'
", "num": null, "type_str": "table", "text": "" }, "TABREF5": { "html": null, "content": "
TaiwaneseSelected Mandarin by SystemAppropriate Mandarin WordRemark
chak/chak\u4fc3(VF) 'urge'\u64e0(VC) 'crowd'
ch\u016bn/ch\u016bn\u7d5e(VC) 'twist'\u9663(Nf) '(a numerary adjunct)'2 times
kah/kah\u548c(Caa) 'and'\u5f97(DE) 'a particle used after a verb'3 times
", "num": null, "type_str": "table", "text": "" }, "TABREF6": { "html": null, "content": "
Taiwanese WordCorresponding Mandarin WordSelected POS by SystemCorrect POS
b\u0113 \u6703/b\u0113-\u0113\uf967\u6703 'be unable to'NbD
\u5edf\u57d5/bi\u014d-ti\u00e2\u207f\u5edf\u524d\u9662 'temple square'NaNc (Na Nc)
\u98df\uf934/chia\u030d h-l\u0101u\uf98e\uf934 'old'NaVH
\u8f49\uf9ba/ch\u014dan-li\u00e1u\u8f49 \u5f8c 'after turning'VHVC Ng
\u725b\u64d4\u7063/G\u00fb-ta\u207f-oan\u725b\u64d4\u7063 '(a place name)'VANc
\u6cd5\uf9d8\u4e0a/hoat-lu\u030d t-si\u014dng \u6cd5\uf9d8\u4e0a 'jural'VCN (Na Ncd)
\u975e\u70ba/hui-\u00fbi\u975e\u70ba 'infamous conduct'AN (A Na)
\u7aae\u5fd7/ki\u00f4ng-ch\u00ec\u7aae\u5fd7 'exhaust the ambition'NaV (VH Na)
\u8f15\u8f15\u4ed4/khin-khin-\u00e1\u8f15\u8f15\u5730 'lightly'NbD (VH DE)
\u751f\u5b50/se\u207f-ki\u00e1\u207f\u751f\u5b69\u5b50 'give birth to a child'NaVA (VH Na)
", "num": null, "type_str": "table", "text": "" }, "TABREF7": { "html": null, "content": "
Left ContextWord and POSCorrect POSRight Contextid
tha\u030d k'read'(VC) pe\u030d h-\u014de-j\u012b
l\u016bn 'discuss' (Na)VE'vernacular writing'(Na) khah-i\u00e2\u207f'better1
than'(VJ) \u2026
si\u016b 'be subjected to'(P)
ch\u00f2 'do'(VC) th\u00e2i-l\u00e2ng 'kill someone'(VA)h\u014dan 'criminal' (VC)Nas\u00ed-h\u00eang 'death penalty'(Na) \u00ea'of'(DE)2
7(Neu) l\u00e2ng 'people'(Na)
l\u00e2ng'people'(Na)chi\u030d t-\u0113 'once' (Nd)Dchia\u030d h-l\u0101u 'old'(VH)3
t\u00f9i 'from' (P)kh\u00ed-th\u00e2u 'beginning'(VH)Nv'waver'(VAC) chi\u016b'then'(D) chin'very'(Dfa) t\u0101ng4
M\u00e1-l\u012b 'a person name'(Nb) \u00ea'of'(DE) l\u0101u-p\u0113'father'(Na)s\u00ed 'dead' (Dfb)VH\u00ea'of'(DE) s\u00ee'time'(Na)10
lo\u030d h-lo\u030d h'down'(VA)
kh\u00f2a\u207f 'look at' (VC)kh\u00ed-kh\u00ed'up'(Nb) VA\u00ea'of'(DE) h\u00e1i-\u00e9ng11
'tide'(Na)
n\u011d-h\u00f3a\u207f 'turned yellow'(VH) \u00e0m-t\u0101m'dim'(VH) \u00ea'of'(DE) l\u014d\u0358 -teng'streetlamp'(Na)chhi\u014d 'shine'(D) VCl\u00f3ng'always'(D) b\u0113'not'(D) hn\u1e21'far'(VH)
", "num": null, "type_str": "table", "text": "" }, "TABREF8": { "html": null, "content": "
ReasonCount Percentage(%)Remark
Selection of Inappropriate Mandarin Word2528.7
Absence of Appropriate Mandarin Word1416.1
Unknown Word1011.5
Personal Name44.6
Propagation Error44.6 Includes an unknown word
Totally5765.5 After discounting the repeat count
", "num": null, "type_str": "table", "text": "" }, "TABREF9": { "html": null, "content": "
GenreNo. of WordsNo. of Tagging ErrorsAccuracy Rate (%)
Prose5494392.2
Novel3723690.3
Drama117893.2
", "num": null, "type_str": "table", "text": "Using Statistical Methods and Mandarin Training Data" } } } }