如何消除countvectorizer中的数字字符 我的代码
cv = CountVectorizer(min_df=50, stop_words='english',max_features = 5000,analyzer='word')
cv_fit_addr=cv.fit_transform(data['Adj_Addr'])
print(cv.get_feature_names())
['01', '02', '03', '04', '05', '06', '07', '08', '09', '10', '100', '1001', '1002', '1003', '1004', '1005', '1008', '101', '1010', '102', '103', '104', '105', '106', '107', '108', '109', '10f', '10th', '11', '1101', '1102', '1103', '1104', '1105', '1106', '1108', '111', '1111', '113', '114', '116', '118', '11f', '11th', '12', '120', '1201', '1202', '1203', '1204', '1206', '1208', '121', '122', '123', '125', '126', '128', '12a', '12f', '12th', '13', '1301', '1302', '1303', '1305', '1308', '132', '133', '135', '138', '139', '13f', '13th', '14', '141', '143', '148', '14f', '14th', '15', '150', '1501', '1502', '1503', '1505', '151', '153', '15f', '15th', '16', '160', '1601', '1602', '1603', '1608', '165', '168', '169', '16f', '16th', '17', '1701', '1702', '1703', '1705', '17f', '17th', '18', '1801', '1803', '181', '182', '183', '188', '18f', '18th', '19', '1901', '1902', '191', '193', '19f', '19th', '1a', '1b', '1f', '1st', '20', '200', '2001', '2003', '201', '202', '203', '204', '205', '206', '208', '20f', '20th', '21', '210', '2101', '2103', '211', '21f', '21st', '22', '220', '2201', '223', '228', '22f', '22nd', '23', '2301', '231', '23f', '23rd', '24', '248', '25', '255', '25f', '25th', '26', '2601', '26f', '26th', '27', '2701', '27f', '28', '28f', '29', '29th', '2a', '2b', '2f', '2g', '2nd', '30', '301', '302', '303', '305', '306', '307', '308', '30f', '31', '311', '32', '33', '338', '34', '35', '36', '37', '370', '38', '388', '39', '392', '3a', '3b', '3f', '3rd', '40', '401', '402', '403', '404', '405', '406', '407', '41', '418', '42', '43', '44', '45', '46', '47', '479', '48', '489', '49', '491', '4a', '4f', '4th', '50', '500', '501', '502', '503', '505', '509', '51', '510', '511', '52', '53', '538', '54', '55', '555', '56', '57', '576', '58', '582', '59', '592', '5a', '5b', '5f', '5th', '60', '601', '602', '603', '605', '607', '608', '609', '61', '610', '611', '62', '625', '63', '64', '65', '66', '67', '68', '681', '69', '6a', '6f', '6th', '70', '701', '702', '703', '704', '705', '706', '707', '71', '712', '72', '73', '74', '75', '76', '760', '77', '777', '778', '78', '788', '79', '7a', '7f', '7th', '80', '800', '801', '802', '803', '804', '805', '806', '807', '808', '81', '810', '82', '83', '833', '838', '84', '852', '87', '88', '883', '89', '8a', '8f', '8th', '901', '902', '903', '904', '905', '906', '907', '908', '909', '912', '92', '93', '94', '95', '979', '98', '99', '9a', '9f', '9th', 'a1', 'a2', 'a3', 'a5', 'aberdeen', 'academic', 'accessories', 'ace', 'admiralty', 'advanced', 'ag', 'aia', 'air', 'airport', 'alexandra', 'alliance', 'allied', 'alpha', 'america', 'ap', 'apartment', 'apparel', 'argyle', 'arrow', 'art', 'ashley', 'asia', 'asset', 'associates', 'atl', 'attn', 'au', 'austin', 'avenue', 'aviation', 'axa', 'b1', 'b2', 'ba', 'bank', 'baptist', 'bay', 'bea', 'bear', 'beauty', 'bel', 'berth', 'best', 'beverly', 'billion', 'bio', 'biotech', 'bldg', 'block', 'blue', 'bo', 'bonham', 'br', 'branch', 'bright', 'broadway', 'bu', 'building', 'bun', 'business', 'c1', 'ca', 'cable', 'cambridge', 'cameron', 'canton', 'capital', 'cargo', 'castle', 'causeway', 'cc', 'cct', 'cent', 'centra'
答案 0 :(得分：2)
由于数值在您的数据中，您应该让您的模型对其进行说明。
为此，您应该在应用NUM之前将数字替换为CountVectorizer之类的通用数字。
如果您的数据位于pandas数据框中，则可以使用NUM替换所有数字，如下所示：
In [7]: pd.DataFrame([['Hello 1'],['01'],["it's 2017"]])
Out[7]:
0
0 Hello 1
1 01
2 it's 2017
In [8]: pd.DataFrame([['Hello 1'],['01'],["it's 2017"]]).replace('\d+', 'NUM', regex=True)
Out[8]:
0
0 Hello NUM
1 NUM
2 it's NUM
然后您可以应用CountVectorizer。之后，print(cv.get_feature_names())输出中的数字将替换为单个NUM。
答案 1 :(得分：1)
您可以将自定义文本df = df.assign(Min_Value=df[Params].min(axis=1),
Min_Param=df[Params].idxmin(axis=1))
df
Params P_01 P_02 P_03 ... P_34 Min_Value Min_Param
ID ...
AA0001 1470.0 NaN NaN ... 1300.0 1200.0 P_14
AA0002 795.0 NaN NaN ... NaN 795.0 P_09
AA0003 1415.0 NaN NaN ... NaN 1200.0 P_14
AA0004 1470.0 NaN NaN ... NaN 1250.0 P_14
AA0005 775.0 NaN NaN ... NaN 598.0 P_08
传递给preprocessor对象，例如：
CountVectorizer答案 2 :(得分：0)
你可以这样过滤掉：
list_1=['01', '02', '03', '04', '05', '06', '07', '08', '09', '10', '100', '1001', '1002', '1003', '1004', '1005', '1008', '101', '1010', '102', '103', '104', '105', '106', '107', '108', '109', '10f', '10th', '11', '1101', '1102', '1103', '1104', '1105', '1106', '1108', '111', '1111', '113', '114', '116', '118', '11f', '11th', '12', '120', '1201', '1202', '1203', '1204', '1206', '1208', '121', '122', '123', '125', '126', '128', '12a', '12f', '12th', '13', '1301', '1302', '1303', '1305', '1308', '132', '133', '135', '138', '139', '13f', '13th', '14', '141', '143', '148', '14f', '14th', '15', '150', '1501', '1502', '1503', '1505', '151', '153', '15f', '15th', '16', '160', '1601', '1602', '1603', '1608', '165', '168', '169', '16f', '16th', '17', '1701', '1702', '1703', '1705', '17f', '17th', '18', '1801', '1803', '181', '182', '183', '188', '18f', '18th', '19', '1901', '1902', '191', '193', '19f', '19th', '1a', '1b', '1f', '1st', '20', '200', '2001', '2003', '201', '202', '203', '204', '205', '206', '208', '20f', '20th', '21', '210', '2101', '2103', '211', '21f', '21st', '22', '220', '2201', '223', '228', '22f', '22nd', '23', '2301', '231', '23f', '23rd', '24', '248', '25', '255', '25f', '25th', '26', '2601', '26f', '26th', '27', '2701', '27f', '28', '28f', '29', '29th', '2a', '2b', '2f', '2g', '2nd', '30', '301', '302', '303', '305', '306', '307', '308', '30f', '31', '311', '32', '33', '338', '34', '35', '36', '37', '370', '38', '388', '39', '392', '3a', '3b', '3f', '3rd', '40', '401', '402', '403', '404', '405', '406', '407', '41', '418', '42', '43', '44', '45', '46', '47', '479', '48', '489', '49', '491', '4a', '4f', '4th', '50', '500', '501', '502', '503', '505', '509', '51', '510', '511', '52', '53', '538', '54', '55', '555', '56', '57', '576', '58', '582', '59', '592', '5a', '5b', '5f', '5th', '60', '601', '602', '603', '605', '607', '608', '609', '61', '610', '611', '62', '625', '63', '64', '65', '66', '67', '68', '681', '69', '6a', '6f', '6th', '70', '701', '702', '703', '704', '705', '706', '707', '71', '712', '72', '73', '74', '75', '76', '760', '77', '777', '778', '78', '788', '79', '7a', '7f', '7th', '80', '800', '801', '802', '803', '804', '805', '806', '807', '808', '81', '810', '82', '83', '833', '838', '84', '852', '87', '88', '883', '89', '8a', '8f', '8th', '901', '902', '903', '904', '905', '906', '907', '908', '909', '912', '92', '93', '94', '95', '979', '98', '99', '9a', '9f', '9th', 'a1', 'a2', 'a3', 'a5', 'aberdeen', 'academic', 'accessories', 'ace', 'admiralty', 'advanced', 'ag', 'aia', 'air', 'airport', 'alexandra', 'alliance', 'allied', 'alpha', 'america', 'ap', 'apartment', 'apparel', 'argyle', 'arrow', 'art', 'ashley', 'asia', 'asset', 'associates', 'atl', 'attn', 'au', 'austin', 'avenue', 'aviation', 'axa', 'b1', 'b2', 'ba', 'bank', 'baptist', 'bay', 'bea', 'bear', 'beauty', 'bel', 'berth', 'best', 'beverly', 'billion', 'bio', 'biotech', 'bldg', 'block', 'blue', 'bo', 'bonham', 'br', 'branch', 'bright', 'broadway', 'bu', 'building', 'bun', 'business', 'c1', 'ca', 'cable', 'cambridge', 'cameron', 'canton', 'capital', 'cargo', 'castle', 'causeway', 'cc', 'cct', 'cent', 'centra']
print([item for item in list_1 if item.isdigit() or item.isalpha()])
输出：
['01', '02', '03', '04', '05', '06', '07', '08', '09', '10', '100', '1001', '1002', '1003', '1004', '1005', '1008', '101', '1010', '102', '103', '104', '105', '106', '107', '108', '109', '11', '1101', '1102', '1103', '1104', '1105', '1106', '1108', '111', '1111', '113', '114', '116', '118', '12', '120', '1201', '1202', '1203', '1204', '1206', '1208', '121', '122', '123', '125', '126', '128', '13', '1301', '1302', '1303', '1305', '1308', '132', '133', '135', '138', '139', '14', '141', '143', '148', '15', '150', '1501', '1502', '1503', '1505', '151', '153', '16', '160', '1601', '1602', '1603', '1608', '165', '168', '169', '17', '1701', '1702', '1703', '1705', '18', '1801', '1803', '181', '182', '183', '188', '19', '1901', '1902', '191', '193', '20', '200', '2001', '2003', '201', '202', '203', '204', '205', '206', '208', '21', '210', '2101', '2103', '211', '22', '220', '2201', '223', '228', '23', '2301', '231', '24', '248', '25', '255', '26', '2601', '27', '2701', '28', '29', '30', '301', '302', '303', '305', '306', '307', '308', '31', '311', '32', '33', '338', '34', '35', '36', '37', '370', '38', '388', '39', '392', '40', '401', '402', '403', '404', '405', '406', '407', '41', '418', '42', '43', '44', '45', '46', '47', '479', '48', '489', '49', '491', '50', '500', '501', '502', '503', '505', '509', '51', '510', '511', '52', '53', '538', '54', '55', '555', '56', '57', '576', '58', '582', '59', '592', '60', '601', '602', '603', '605', '607', '608', '609', '61', '610', '611', '62', '625', '63', '64', '65', '66', '67', '68', '681', '69', '70', '701', '702', '703', '704', '705', '706', '707', '71', '712', '72', '73', '74', '75', '76', '760', '77', '777', '778', '78', '788', '79', '80', '800', '801', '802', '803', '804', '805', '806', '807', '808', '81', '810', '82', '83', '833', '838', '84', '852', '87', '88', '883', '89', '901', '902', '903', '904', '905', '906', '907', '908', '909', '912', '92', '93', '94', '95', '979', '98', '99', 'aberdeen', 'academic', 'accessories', 'ace', 'admiralty', 'advanced', 'ag', 'aia', 'air', 'airport', 'alexandra', 'alliance', 'allied', 'alpha', 'america', 'ap', 'apartment', 'apparel', 'argyle', 'arrow', 'art', 'ashley', 'asia', 'asset', 'associates', 'atl', 'attn', 'au', 'austin', 'avenue', 'aviation', 'axa', 'ba', 'bank', 'baptist', 'bay', 'bea', 'bear', 'beauty', 'bel', 'berth', 'best', 'beverly', 'billion', 'bio', 'biotech', 'bldg', 'block', 'blue', 'bo', 'bonham', 'br', 'branch', 'bright', 'broadway', 'bu', 'building', 'bun', 'business', 'ca', 'cable', 'cambridge', 'cameron', 'canton', 'capital', 'cargo', 'castle', 'causeway', 'cc', 'cct', 'cent', 'centra']