mstz commited on
Commit
4db99e8
1 Parent(s): 8d066cd

Upload german.py

Browse files
Files changed (1) hide show
  1. german.py +95 -345
german.py CHANGED
@@ -1,4 +1,4 @@
1
- """Breast Dataset"""
2
 
3
  from typing import List
4
 
@@ -54,13 +54,83 @@ _BASE_FEATURE_NAMES = [
54
  "is_foreign",
55
  "loan_granted"
56
  ]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
57
 
58
- DESCRIPTION = "Breast dataset for cancer prediction."
59
  _HOMEPAGE = "https://archive.ics.uci.edu/ml/datasets/Statlog+%28German+Credit+Data%29"
60
  _URLS = ("https://archive.ics.uci.edu/ml/datasets/Statlog+%28German+Credit+Data%29")
61
- _CITATION = """
62
-
63
- """
64
 
65
  # Dataset info
66
  urls_per_split = {
@@ -101,19 +171,19 @@ features_types_per_config = {
101
  features_per_config = {k: datasets.Features(features_types_per_config[k]) for k in features_types_per_config}
102
 
103
 
104
- class BreastConfig(datasets.BuilderConfig):
105
  def __init__(self, **kwargs):
106
- super(BreastConfig, self).__init__(version=VERSION, **kwargs)
107
  self.features = features_per_config[kwargs["name"]]
108
 
109
 
110
- class Breast(datasets.GeneratorBasedBuilder):
111
  # dataset versions
112
  DEFAULT_CONFIG = "loan"
113
  BUILDER_CONFIGS = [
114
- BreastConfig(name="encoding",
115
  description="Encoding dictionaries for discrete features."),
116
- BreastConfig(name="loan",
117
  description="Binary classification of loan approval."),
118
  ]
119
 
@@ -135,8 +205,12 @@ class Breast(datasets.GeneratorBasedBuilder):
135
  ]
136
 
137
  def _generate_examples(self, filepath: str):
138
- if self.config.name == "encoding":
139
- data = self.encoding_dictionaries()
 
 
 
 
140
  else:
141
  data = pandas.read_csv(filepath, sep=" ", header=None)
142
  data.columns=_ORIGINAL_FEATURE_NAMES
@@ -147,54 +221,11 @@ class Breast(datasets.GeneratorBasedBuilder):
147
 
148
  yield row_id, data_row
149
 
150
- def encoding_dictionaries(self):
151
- dictionaries = list([
152
- self.checking_account_status_encode_dic(),
153
- self.credit_status_encode_dic(),
154
- self.current_savings_encode_dic(),
155
- self.employed_since_encode_dic(),
156
- self.guarantors_encode_dic(),
157
- self.has_registered_phone_number_encode_dic(),
158
- self.housing_status_encode_dic(),
159
- self.installment_plans_encode_dic(),
160
- self.is_foreign_encode_dic(),
161
- self.job_status_encode_dic(),
162
- self.marital_status_encode_dic(),
163
- self.sex_encode_dic(),
164
- ])
165
- data = [[(feature, value, code) for value, code in d.items()]
166
- for feature, d in zip(["checking_account_status",
167
- "credit_status",
168
- "current_savings",
169
- "employed_since",
170
- "guarantors",
171
- "has_registered_phone_number",
172
- "housing_status",
173
- "installment_plans",
174
- "is_foreign",
175
- "job_status",
176
- "marital_status",
177
- "sex"],
178
- dictionaries)]
179
- full_data = list()
180
- for d in data:
181
- full_data += d
182
- data = pandas.DataFrame(full_data,
183
- columns=["feature", "original_value", "encoded_value"])
184
-
185
- return data
186
-
187
- def preprocess(self, data: pandas.DataFrame, config: str = "cancer") -> pandas.DataFrame:
188
- data.loc[:, "checking_account_status"] = data.checking_account_status.apply(self.encode_checking_account_status)
189
- data.loc[:, "credit_status"] = data.credit_status.apply(self.encode_credit_status)
190
- data.loc[:, "current_savings"] = data.current_savings.apply(self.encode_current_savings)
191
- data.loc[:, "employed_since"] = data.employed_since.apply(self.encode_employed_since)
192
- data.loc[:, "guarantors"] = data.guarantors.apply(self.encode_guarantors)
193
- data.loc[:, "has_registered_phone_number"] = data.has_registered_phone_number.apply(self.encode_has_registered_phone_number)
194
- data.loc[:, "housing_status"] = data.housing_status.apply(self.encode_housing_status)
195
- data.loc[:, "installment_plans"] = data.installment_plans.apply(self.encode_installment_plans)
196
- data.loc[:, "is_foreign"] = data.is_foreign.apply(self.encode_is_foreign)
197
- data.loc[:, "job_status"] = data.job_status.apply(self.encode_job_status)
198
  data.loc[:, "marital_status"] = data.personal_status_and_sex.apply(self.encode_marital_status)
199
  data.loc[:, "sex"] = data.personal_status_and_sex.apply(self.encode_sex)
200
  data.loc[:, "loan_purpose"] = data.loan_purpose.apply(self.encode_loan_purpose)
@@ -202,274 +233,11 @@ class Breast(datasets.GeneratorBasedBuilder):
202
 
203
  data.drop("personal_status_and_sex", axis="columns", inplace=True)
204
 
205
- print(data.sex.unique())
206
-
207
  data = data[_BASE_FEATURE_NAMES]
208
 
209
- if config == "loan":
210
- return data
211
- else:
212
- raise ValueError(f"Unknown config: {config}")
213
-
214
- def encode_checking_account_status(self, status):
215
- return self.checking_account_status_encode_dic()[status]
216
-
217
- def checking_account_status_encode_dic(self):
218
- return {
219
- "A14": 0,
220
- "A11": 1,
221
- "A12": 2,
222
- "A13": 3,
223
- }
224
-
225
- def decode_checking_account_status(self, code):
226
- return self.checking_account_status_decode_dic()[code]
227
-
228
- def checking_account_status_decode_dic(self):
229
- return {
230
- 0: "A14",
231
- 1: "A11",
232
- 2: "A12",
233
- 3: "A13",
234
- }
235
-
236
- def encode_credit_status(self, status):
237
- return self.credit_status_encode_dic()[status]
238
-
239
- def credit_status_encode_dic(self):
240
- return {
241
- "A30": 0,
242
- "A31": 1,
243
- "A32": 2,
244
- "A33": 3,
245
- "A34": 4,
246
- }
247
-
248
- def decode_credit_status(self, code):
249
- return self.credit_status_decode_dic()[code]
250
-
251
- def credit_status_decode_dic(self):
252
- return {
253
- 0: "A30",
254
- 1: "A31",
255
- 2: "A32",
256
- 3: "A33",
257
- 4: "A34",
258
- }
259
-
260
- def encode_current_savings(self, status):
261
- return self.current_savings_encode_dic()[status]
262
-
263
- def current_savings_encode_dic(self):
264
- return {
265
- "A65": 0,
266
- "A61": 1,
267
- "A62": 2,
268
- "A63": 3,
269
- "A64": 4,
270
- }
271
-
272
- def decode_current_savings(self, code):
273
- return self.current_savings_decode_dic()[code]
274
-
275
- def current_savings_decode_dic(self):
276
- return {
277
- 0: "A65",
278
- 1: "A61",
279
- 2: "A62",
280
- 3: "A63",
281
- 4: "A64",
282
- }
283
-
284
- def encode_employed_since(self, status):
285
- return self.employed_since_encode_dic()[status]
286
-
287
- def employed_since_encode_dic(self):
288
- return {
289
- "A71": 0,
290
- "A72": 1,
291
- "A73": 2,
292
- "A74": 3,
293
- "A75": 4,
294
- }
295
-
296
- def decode_employed_since(self, code):
297
- return self.employed_since_decode_dic()[code]
298
-
299
- def employed_since_decode_dic(self):
300
- return {
301
- 0: "A71",
302
- 1: "A72",
303
- 2: "A73",
304
- 3: "A74",
305
- 4: "A75",
306
- }
307
-
308
- def encode_sex(self, status):
309
- return self.sex_encode_dic()[status]
310
-
311
- def sex_encode_dic(self):
312
- return {
313
- "A91": 0,
314
- "A93": 0,
315
- "A94": 0,
316
- "A92": 1,
317
- "A95": 1,
318
- }
319
-
320
- def decode_sex(self, code):
321
- return self.sex_decode_dic()[code]
322
-
323
- def sex_decode_dic(self):
324
- return {
325
- 0: "A91",
326
- 1: "A92",
327
- }
328
-
329
- def encode_marital_status(self, status):
330
- return self.marital_status_encode_dic()[status]
331
-
332
- def marital_status_encode_dic(self):
333
- return {
334
- "A91": 0,
335
- "A92": 0,
336
- "A93": 1,
337
- "A94": 2,
338
- "A95": 1,
339
- }
340
-
341
- def decode_marital_status(self, code):
342
- return self.marital_status_decode_dic()[code]
343
-
344
- def marital_status_decode_dic(self):
345
- return {
346
- 0: "A91",
347
- 1: "A93",
348
- 2: "A94",
349
- }
350
-
351
- def encode_guarantors(self, status):
352
- return self.guarantors_encode_dic()[status]
353
-
354
- def guarantors_encode_dic(self):
355
- return {
356
- "A101": 0,
357
- "A102": 1,
358
- "A103": 2,
359
- }
360
-
361
- def decode_guarantors(self, code):
362
- return self.guarantors_decode_dic()[code]
363
-
364
- def guarantors_decode_dic(self):
365
- return {
366
- 0: "A101",
367
- 1: "A102",
368
- 2: "A103",
369
- }
370
-
371
- def encode_installment_plans(self, status):
372
- return self.installment_plans_encode_dic()[status]
373
-
374
- def installment_plans_encode_dic(self):
375
- return {
376
- "A141": 0,
377
- "A142": 1,
378
- "A143": 2,
379
- }
380
-
381
- def decode_installment_plans(self, code):
382
- return self.installment_plans_decode_dic()[code]
383
-
384
- def installment_plans_decode_dic(self):
385
- return {
386
- 0: "A141",
387
- 1: "A142",
388
- 2: "A143",
389
- }
390
-
391
- def encode_housing_status(self, status):
392
- return self.housing_status_encode_dic()[status]
393
-
394
- def housing_status_encode_dic(self):
395
- return {
396
- "A153": 0,
397
- "A151": 1,
398
- "A152": 2,
399
- }
400
-
401
- def decode_housing_status(self, code):
402
- return self.housing_status_decode_dic()[code]
403
-
404
- def housing_status_decode_dic(self):
405
- return {
406
- 0: "A153",
407
- 1: "A151",
408
- 2: "A152",
409
- }
410
-
411
- def encode_job_status(self, status):
412
- return self.job_status_encode_dic()[status]
413
-
414
- def job_status_encode_dic(self):
415
- return {
416
- "A171": 0,
417
- "A172": 1,
418
- "A173": 2,
419
- "A174": 3,
420
- }
421
-
422
- def decode_job_status(self, code):
423
- return self.job_status_decode_dic()[code]
424
-
425
- def job_status_decode_dic(self):
426
- return {
427
- 0: "A171",
428
- 1: "A172",
429
- 2: "A173",
430
- 3: "A174",
431
- }
432
-
433
- def encode_has_registered_phone_number(self, status):
434
- return self.has_registered_phone_number_encode_dic()[status]
435
-
436
- def has_registered_phone_number_encode_dic(self):
437
- return {
438
- "A191": 0,
439
- "A192": 1,
440
- }
441
-
442
- def decode_has_registered_phone_number(self, code):
443
- return self.has_registered_phone_number_decode_dic()[code]
444
-
445
- def has_registered_phone_number_decode_dic(self):
446
- return {
447
- 0: "A191",
448
- 1: "A192",
449
- }
450
-
451
- def encode_is_foreign(self, status):
452
- return self.is_foreign_encode_dic()[status]
453
-
454
- def is_foreign_encode_dic(self):
455
- return {
456
- "A201": 0,
457
- "A202": 1,
458
- }
459
-
460
- def decode_is_foreign(self, code):
461
- return self.is_foreign_decode_dic()[code]
462
-
463
- def is_foreign_decode_dic(self):
464
- return {
465
- 0: "A201",
466
- 1: "A202",
467
- }
468
-
469
- def encode_loan_purpose(self, status):
470
- return self.loan_purpose_encode_dic()[status]
471
-
472
- def loan_purpose_encode_dic(self):
473
  return {
474
  "A40": "new car",
475
  "A41": "used car",
@@ -482,22 +250,4 @@ class Breast(datasets.GeneratorBasedBuilder):
482
  "A48": "retraining",
483
  "A49": "business",
484
  "A410": "others"
485
- }
486
-
487
- def decode_loan_purpose(self, code):
488
- return self.loan_purpose_decode_dic()[code]
489
-
490
- def loan_purpose_decode_dic(self):
491
- return {
492
- "new car": "A40",
493
- "used car": "A41",
494
- "furniture/equipment": "A42",
495
- "radio/television": "A43",
496
- "domestic appliances": "A44",
497
- "repairs": "A45",
498
- "education": "A46",
499
- "vacation ": "A47",
500
- "retraining": "A48",
501
- "business": "A49",
502
- "others": "A410"
503
- }
 
1
+ """German Dataset"""
2
 
3
  from typing import List
4
 
 
54
  "is_foreign",
55
  "loan_granted"
56
  ]
57
+ _ENCODING_DICS = {
58
+ "is_foreign": {
59
+ "A201": 0,
60
+ "A202": 1
61
+ },
62
+ "has_registered_phone_number": {
63
+ "A191": 0,
64
+ "A192": 1
65
+ },
66
+ "job_status": {
67
+ "A171": 0,
68
+ "A172": 1,
69
+ "A173": 2,
70
+ "A174": 3
71
+ },
72
+ "housing_status": {
73
+ "A153": 0,
74
+ "A151": 1,
75
+ "A152": 2
76
+ },
77
+ "installment_plans": {
78
+ "A141": 0,
79
+ "A142": 1,
80
+ "A143": 2
81
+ },
82
+ "guarantors": {
83
+ "A101": 0,
84
+ "A102": 1,
85
+ "A103": 2
86
+ },
87
+ "marital_status": {
88
+ "A91": 0,
89
+ "A92": 0,
90
+ "A93": 1,
91
+ "A94": 2,
92
+ "A95": 1,
93
+ },
94
+ "sex": {
95
+ "A91": 0,
96
+ "A93": 0,
97
+ "A94": 0,
98
+ "A92": 1,
99
+ "A95": 1,
100
+ },
101
+ "employed_since": {
102
+ "A71": 0,
103
+ "A72": 1,
104
+ "A73": 2,
105
+ "A74": 3,
106
+ "A75": 4,
107
+ },
108
+ "current_savings": {
109
+ "A65": 0,
110
+ "A61": 1,
111
+ "A62": 2,
112
+ "A63": 3,
113
+ "A64": 4,
114
+ },
115
+ "credit_status": {
116
+ "A30": 0,
117
+ "A31": 1,
118
+ "A32": 2,
119
+ "A33": 3,
120
+ "A34": 4,
121
+ },
122
+ "checking_account_status": {
123
+ "A14": 0,
124
+ "A11": 1,
125
+ "A12": 2,
126
+ "A13": 3,
127
+ }
128
+ }
129
 
130
+ DESCRIPTION = "German dataset for cancer prediction."
131
  _HOMEPAGE = "https://archive.ics.uci.edu/ml/datasets/Statlog+%28German+Credit+Data%29"
132
  _URLS = ("https://archive.ics.uci.edu/ml/datasets/Statlog+%28German+Credit+Data%29")
133
+ _CITATION = """"""
 
 
134
 
135
  # Dataset info
136
  urls_per_split = {
 
171
  features_per_config = {k: datasets.Features(features_types_per_config[k]) for k in features_types_per_config}
172
 
173
 
174
+ class GermanConfig(datasets.BuilderConfig):
175
  def __init__(self, **kwargs):
176
+ super(GermanConfig, self).__init__(version=VERSION, **kwargs)
177
  self.features = features_per_config[kwargs["name"]]
178
 
179
 
180
+ class German(datasets.GeneratorBasedBuilder):
181
  # dataset versions
182
  DEFAULT_CONFIG = "loan"
183
  BUILDER_CONFIGS = [
184
+ GermanConfig(name="encoding",
185
  description="Encoding dictionaries for discrete features."),
186
+ GermanConfig(name="loan",
187
  description="Binary classification of loan approval."),
188
  ]
189
 
 
205
  ]
206
 
207
  def _generate_examples(self, filepath: str):
208
+ if self.config.name not in self.features_per_config:
209
+ raise ValueError(f"Unknown config: {self.config.name}")
210
+
211
+ elif self.config.name == "encoding":
212
+ data = self.encoding_dics()
213
+
214
  else:
215
  data = pandas.read_csv(filepath, sep=" ", header=None)
216
  data.columns=_ORIGINAL_FEATURE_NAMES
 
221
 
222
  yield row_id, data_row
223
 
224
+ def preprocess(self, data: pandas.DataFrame, config: str = "loan") -> pandas.DataFrame:
225
+ for feature in _ENCODING_DICS:
226
+ if feature not in ["marital_status", "sex"]:
227
+ encoding_function = partial(self.encode, feature)
228
+ data.loc[:, feature] = data[feature].apply(encoding_function)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
229
  data.loc[:, "marital_status"] = data.personal_status_and_sex.apply(self.encode_marital_status)
230
  data.loc[:, "sex"] = data.personal_status_and_sex.apply(self.encode_sex)
231
  data.loc[:, "loan_purpose"] = data.loan_purpose.apply(self.encode_loan_purpose)
 
233
 
234
  data.drop("personal_status_and_sex", axis="columns", inplace=True)
235
 
 
 
236
  data = data[_BASE_FEATURE_NAMES]
237
 
238
+ return data
239
+
240
+ def encode_loan_purpose(self, code):
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
241
  return {
242
  "A40": "new car",
243
  "A41": "used car",
 
250
  "A48": "retraining",
251
  "A49": "business",
252
  "A410": "others"
253
+ }[code]