/
uci_038_tv_news_channel_commercial_detection.py
199 lines (182 loc) · 7.37 KB
/
uci_038_tv_news_channel_commercial_detection.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import urllib.request
import io
import zipfile
import sklearn.datasets
import pandas # install pandas by "pip install pandas", or install Anaconda distribution (https://www.anaconda.com/)
# Warning: the data processing techniques shown below are just for concept explanation, which are not best-proctices
# data set repository
# https://archive.ics.uci.edu/ml/datasets/TV+News+Channel+Commercial+Detection+Dataset
# if the file is on your local device, change url_data_train into local file path, e.g., 'D:\local_file.data'
url_data_train = 'https://archive.ics.uci.edu/ml/machine-learning-databases/00326/TV_News_Channel_Commercial_Detection_Dataset.zip'
def download_file(url):
resp = urllib.request.urlopen(url)
if resp.status != 200:
resp.close()
raise ValueError('Error: {0}'.format(resp.reason))
print('\rStarted', end = '\r')
content_length = resp.getheader('Content-Length')
if content_length is None:
content_length = '(total: unknown)'
else:
content_length = int(content_length)
if content_length < 1024:
content_length_str = '(total %.0f Bytes)' % content_length
elif content_length < 1024 * 1024:
content_length_str = '(total %.0f KB)' % (content_length / 1024)
else:
content_length_str = '(total %.1f MB)' % (content_length / 1024 / 1024)
total = bytes()
while not resp.isclosed():
total += resp.read(10 * 1024)
if len(total) < 1024:
print(('\rDownloaded: %.0f Bytes ' % len(total)) + content_length_str + ' ', end = '\r')
if len(total) < 1024 * 1024:
print(('\rDownloaded: %.0f KB ' % (len(total) / 1024)) + content_length_str + ' ', end = '\r')
else:
print(('\rDownloaded: %.1f MB ' % (len(total) / 1024 / 1024)) + content_length_str + ' ', end = '\r')
print()
return io.BytesIO(total)
# download data from UCI Machine Learning Repository
data_train = download_file(url_data_train) if url_data_train.startswith('http') else url_data_train
audio_word_columns = ['word_' + str(i + 1).zfill(4) for i in range(4000)]
columns = [
'Shot_Length',
'Motion_Distribution_Mean',
'Motion_Distribution_Variance',
'Frame_Difference_Distribution_Mean',
'Frame_Difference_Distribution_Variance',
'Short_time_energy_Mean',
'Short_time_energy_Variance',
'ZCR_Mean',
'ZCR_Variance',
'Spectral_Centroid_Mean',
'Spectral_Centroid_Variance',
'Spectral_Roll_off_Mean',
'Spectral_Roll_off_Variance',
'Spectral_Flux_Mean',
'Spectral_Flux_Variance',
'Fundamental_Frequency_Mean',
'Fundamental_Frequency_Variance',
'Motion_Distribution_01',
'Motion_Distribution_02',
'Motion_Distribution_03',
'Motion_Distribution_04',
'Motion_Distribution_05',
'Motion_Distribution_06',
'Motion_Distribution_07',
'Motion_Distribution_08',
'Motion_Distribution_09',
'Motion_Distribution_10',
'Motion_Distribution_11',
'Motion_Distribution_12',
'Motion_Distribution_13',
'Motion_Distribution_14',
'Motion_Distribution_15',
'Motion_Distribution_16',
'Motion_Distribution_17',
'Motion_Distribution_18',
'Motion_Distribution_19',
'Motion_Distribution_20',
'Motion_Distribution_21',
'Motion_Distribution_22',
'Motion_Distribution_23',
'Motion_Distribution_24',
'Motion_Distribution_25',
'Motion_Distribution_26',
'Motion_Distribution_27',
'Motion_Distribution_28',
'Motion_Distribution_29',
'Motion_Distribution_30',
'Motion_Distribution_31',
'Motion_Distribution_32',
'Motion_Distribution_33',
'Motion_Distribution_34',
'Motion_Distribution_35',
'Motion_Distribution_36',
'Motion_Distribution_37',
'Motion_Distribution_38',
'Motion_Distribution_39',
'Motion_Distribution_40',
'Motion_Distribution_41',
'Frame_Difference_Distribution_01',
'Frame_Difference_Distribution_02',
'Frame_Difference_Distribution_03',
'Frame_Difference_Distribution_04',
'Frame_Difference_Distribution_05',
'Frame_Difference_Distribution_06',
'Frame_Difference_Distribution_07',
'Frame_Difference_Distribution_08',
'Frame_Difference_Distribution_09',
'Frame_Difference_Distribution_10',
'Frame_Difference_Distribution_11',
'Frame_Difference_Distribution_12',
'Frame_Difference_Distribution_13',
'Frame_Difference_Distribution_14',
'Frame_Difference_Distribution_15',
'Frame_Difference_Distribution_16',
'Frame_Difference_Distribution_17',
'Frame_Difference_Distribution_18',
'Frame_Difference_Distribution_19',
'Frame_Difference_Distribution_20',
'Frame_Difference_Distribution_21',
'Frame_Difference_Distribution_22',
'Frame_Difference_Distribution_23',
'Frame_Difference_Distribution_24',
'Frame_Difference_Distribution_25',
'Frame_Difference_Distribution_26',
'Frame_Difference_Distribution_27',
'Frame_Difference_Distribution_28',
'Frame_Difference_Distribution_29',
'Frame_Difference_Distribution_30',
'Frame_Difference_Distribution_31',
'Frame_Difference_Distribution_32',
'Frame_Difference_Distribution_33',
'Text_area_distribution_Mean_01',
'Text_area_distribution_Mean_02',
'Text_area_distribution_Mean_03',
'Text_area_distribution_Mean_04',
'Text_area_distribution_Mean_05',
'Text_area_distribution_Mean_06',
'Text_area_distribution_Mean_07',
'Text_area_distribution_Mean_08',
'Text_area_distribution_Mean_09',
'Text_area_distribution_Mean_10',
'Text_area_distribution_Mean_11',
'Text_area_distribution_Mean_12',
'Text_area_distribution_Mean_13',
'Text_area_distribution_Mean_14',
'Text_area_distribution_Mean_15',
'Text_area_distribution_Mean_16',
'Text_area_distribution_variance_01',
'Text_area_distribution_variance_02',
'Text_area_distribution_variance_03',
'Text_area_distribution_variance_04',
'Text_area_distribution_variance_05',
'Text_area_distribution_variance_06',
'Text_area_distribution_variance_07',
'Text_area_distribution_variance_08',
'Text_area_distribution_variance_09',
'Text_area_distribution_variance_10',
'Text_area_distribution_variance_11',
'Text_area_distribution_variance_12',
'Text_area_distribution_variance_13',
'Text_area_distribution_variance_14',
'Text_area_distribution_variance_15',
'Text_area_distribution_variance_16'] + audio_word_columns + ['Edge_change_Ratio_Mean', 'Edge_change_Ratio_Variance']
# unzip the downloaded file, and get data files
with zipfile.ZipFile(data_train) as myzip:
with myzip.open('CNN.txt') as myfile:
df_x, df_y = sklearn.datasets.load_svmlight_file(myfile)
# convert libsvm format files into pandas dataframes
df_x = pandas.DataFrame(df_x.todense(), columns = columns)
df_y = pandas.DataFrame(df_y, columns = ['label'])
# drop the 4000 audio word columns, because there are too many of them
#df_x = df_x.drop(audio_word_columns, axis = 1)
# merge y with x variables
df_total = df_y.merge(df_x, how = 'left', left_index = True, right_index = True)
# convert label values (originally 1 and -1) into 0 and 1
df_total['label'] = df_total['label'].apply(lambda x: 1 if x == 1 else 0)
# save the dataframe as CSV file, you can zip it, upload it to t1modeler.com, and build a model
df_total.to_csv('uci_038_tv_news_channel_commercial_detection.csv', index = False)