-
Notifications
You must be signed in to change notification settings - Fork 1
/
naive bayes.py
301 lines (257 loc) · 10.2 KB
/
naive bayes.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
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
from __future__ import division
#
# Naive Bayes.py
# Author: Divyakumar Menghani
# Description: This code reads the dataset into pandas dataframes, builds a Naive Bayes Classifier, predicts labels for a subset of data. It also calculates metrics such as precision/recall/accuracy and F-Score after classification. The output is dumped in pickle files which are used later for visualization
#
import pickle
import sklearn
from sklearn.naive_bayes import *
import pandas as pd
import numpy as np
from sklearn import *
import os
from sklearn.metrics import *
from sklearn import metrics, preprocessing
from sklearn import svm, naive_bayes, neighbors, tree
#
# Function: createPickle()
# Description: This function will create a pickle file.
# Input: data structure that you want to pickle
# Output: a pickle file for the data structure. The file is stored in the
# same path the code is running from
#
def createPickle(data, filename):
with open(filename, 'wb') as f:
pickle.dump(data, f)
print "Pickled", filename
# Global constants for this code
print "Setting constants..."
TRAINING_LINE_NUMBER = 8000000 # Number of lines to be read from input files
# List of years for training and testing
YEARS = ['2001', '2002', '2003', '2004', '2005', '2006', '2007', '2008']
INPUT_FILE_PATH = "/home/dmenghani/python/" # Unix path
# INPUT_FILE_PATH = "C:\\data\\airline\\" # Windows path
SKIP_FIRST_LINE = True # To skip the first line, as its the header
# Creating the master data frame from all years.
master = []
print "Reading into Pandas frame..."
try:
for year in YEARS:
path = os.path.join(INPUT_FILE_PATH, '%d.csv' % int(year))
print "\n", path
dfPart = pd.read_csv(
path, nrows=TRAINING_LINE_NUMBER, skiprows=0, usecols=[
u'Year',
u'Month',
u'DayofMonth',
u'DayOfWeek',
u'UniqueCarrier',
u'DepTime',
u'TailNum',
u'Origin',
u'Dest',
u'DepDelay',
# u'ArrDelay',
u'Cancelled',
# u'ArrTime',
# u'ArrDelay',
# u'Distance'
])
print len(dfPart)
# Removing cancelled flights from each year
dfPart = dfPart[dfPart['Cancelled'] == 0]
rows = np.random.choice(
np.random.permutation(dfPart.index.values), len(dfPart) // 3, replace=False) # 33% sampling of training data
print rows
sampled_dfPart = dfPart.ix[rows]
sampled_dfPart = dfPart
master.append(sampled_dfPart)
print
except Exception as e:
print "Supplemental Data Import failed", e
# Building the master frame by concating it for all years
dfMaster = pd.concat(master, ignore_index=True)
master = []
dfPart = []
print "Total length - ", len(dfMaster)
del dfMaster['Cancelled'] # Column not needed
dfMaster.fillna(0, inplace=True)
# Converting to appropriate datatypes for numeric cols.
dfMaster['Year'] = dfMaster['Year'].astype('int')
dfMaster['Month'] = dfMaster['Month'].astype('int')
dfMaster['DayofMonth'] = dfMaster['DayofMonth'].astype('int')
dfMaster['DayOfWeek'] = dfMaster['DayOfWeek'].astype('int')
dfMaster['DepTime'] = dfMaster['DepTime'].astype('int')
dfMaster['DepDelay'] = dfMaster['DepDelay'].astype('int')
df = dfMaster
# Since we dont have a classification label in the data, we are creating
# one. Threshold of 5mins was chosen.
print "Calculating classification label..."
df['label'] = 0
df.label[df.DepDelay >= 5] = 1
df.label[df.DepDelay < 5] = 0
print "Actual delayed flights -", np.sum(dfMaster['label']) / len(dfMaster['label'])
del df['DepDelay']
print "Dataframe shape - ", df.shape
print "Columns -", df.columns
# Converting categorical data to numeric for cols - TailNum,
# UniqueCarrier, Dest, Origin
print "Converting categorical data to numeric..."
for col in set(df.columns):
if df[col].dtype == np.dtype('object'):
print "Converting...", col
if col == 'TailNum':
s = np.unique(df[col].values)
TailNum = pd.Series([x[0] for x in enumerate(s)], index=s)
if col == 'UniqueCarrier':
s = np.unique(df[col].values)
UniqueCarrier = pd.Series([x[0] for x in enumerate(s)], index=s)
if col == 'Dest':
s = np.unique(df[col].values)
Dest = pd.Series([x[0] for x in enumerate(s)], index=s)
if col == 'Origin':
s = np.unique(df[col].values)
Origin = pd.Series([x[0] for x in enumerate(s)], index=s)
# Creating Pickle files for the list containing key-value pairs
createPickle(Dest, 'Dest_2008.pkl')
createPickle(Origin, 'Origin_2008.pkl')
createPickle(UniqueCarrier, 'UniqueCarrier_2008.pkl')
createPickle(TailNum, 'TailNum_2008.pkl')
print "Pickle completed."
#
# Function: getTailNum()
# Description: This function will convert the input categorical value to corresponding numeric key.
# Input: categorical value you want to convert
# Output: a numeric value corresponding to the value passed. It uses the list created previously for lookup.
#
def getTailNum(inTailNum):
out = []
for x, y in inTailNum.iteritems():
out.append(TailNum.get_value(y))
return out
#
# Function: getDest()
# Description: This function will convert the input categorical value to corresponding numeric key.
# Input: categorical value you want to convert
# Output: a numeric value corresponding to the value passed. It uses the list created previously for lookup.
#
def getDest(inDest):
out = []
for x, y in inDest.iteritems():
out.append(Dest.get_value(y))
return out
#
# Function: getOrigin()
# Description: This function will convert the input categorical value to corresponding numeric key.
# Input: categorical value you want to convert
# Output: a numeric value corresponding to the value passed. It uses the list created previously for lookup.
#
def getOrigin(inOrign):
out = []
for x, y in inOrign.iteritems():
out.append(Origin.get_value(y))
return out
#
# Function: getCarrier()
# Description: This function will convert the input categorical value to corresponding numeric key.
# Input: categorical value you want to convert
# Output: a numeric value corresponding to the value passed. It uses the list created previously for lookup.
#
def getCarrier(inCarrier):
out = []
for x, y in inCarrier.iteritems():
out.append(UniqueCarrier.get_value(y))
return out
# Converting TailNum
df['TailNum'] = getTailNum(df['TailNum'])
print "TailNum completed."
# Converting UniqueCarrier
df['UniqueCarrier'] = getCarrier(df['UniqueCarrier'])
print "UniqueCarrier completed."
# Converting Dest
df['Dest'] = getDest(df['Dest'])
print "Dest completed."
# Converting Origin
df['Origin'] = getOrigin(df['Origin'])
print "Origin completed."
print "Conversion to numeric completed."
# Building classifier
print "Begin cross validation..."
# Choosing features for classifier
features = df.columns[0:9]
# Creating lists for storing results for cross validation.
accuracy = {}
results = {}
matrix = {}
prec = {}
recall = {}
for year in YEARS:
print "Testing on - ", year
train = df[df['Year'] != int(year)] # Test on 1year, train on other 7years
test = df[df['Year'] == int(year)]
# test = test[test['Origin'].isin([Origin['OAK'], Origin['SFO']])]
print len(train), len(test)
rows = np.random.choice(np.random.permutation(
test.index.values), len(test) // 2, replace=False) # 50% sampling of test data to avoid memory errors faced.
# print rows
sampled_test = test.ix[rows]
sampled_test = test
# Putting the last column of Training data into a list
trainTargets = np.array(train['label']).astype(int)
# Putting the last column of Testing data into a list
testTargets = np.array(sampled_test['label']).astype(int)
print "Train length - ", len(train), "Test length - ", len(sampled_test)
print train['Year']
print test['Year']
print "Model fitting and prediction started..."
# Building the classifier and fitting the train data
gnb = GaussianNB()
y_gnb = gnb.fit(train[features], trainTargets).predict(
sampled_test[features])
# Storing results in a new colum in the dataframe.
sampled_test['pred_label'] = y_gnb
print "Classification completed."
# Creating pickle files with the classifier and the results of classifier
createPickle(gnb, INPUT_FILE_PATH + "classifier_" + year + ".pkl")
createPickle(y_gnb, INPUT_FILE_PATH + "label_" + year + ".pkl")
sampled_test.to_csv(
INPUT_FILE_PATH + "\dfTest" + year + ".csv", index=False)
# Calculating metrics using sklearn metrics functions
print "\nCalculating metrcs..."
accuracy[int(year)] = accuracy_score(sampled_test['label'], y_gnb)
print "Accuracy score - ", accuracy[int(year)]
prec[int(year)] = precision_score(
sampled_test['label'], y_gnb, average='micro')
print "Precision Score - ", prec[int(year)]
recall[int(year)] = recall_score(
sampled_test['label'], y_gnb, average='micro')
print "Recall Score - ", recall[int(year)]
print "Confusion matrix"
matrix[int(year)] = metrics.confusion_matrix(
sampled_test['label'], y_gnb)
print matrix[int(year)]
results[int(year)] = precision_recall_fscore_support(
sampled_test['label'], y_gnb, average='micro')
print "Precision, recall, F-Score, Support - ", results[int(year)]
print "Classification report"
print classification_report(np.array(sampled_test['label']), y_gnb,
target_names=target_names)
print
train = []
test = []
print "Accuracy\n", accuracy
print "\nPrecision\n", prec
print "\nRecall\n", recall
print "\nMetrics\n", results
print "\nMatrix\n", matrix
# Finding mean of metrics
print "\nMean Cross validation Precision score", np.mean(pd.Series(prec))
print "\nMean Cross validation Recall score", np.mean(pd.Series(recall))
print "\nMean Cross validation Accuracy score", np.mean(pd.Series(accuracy))
# Pickling results
print "\nPickling stuff..."
createPickle(accuracy, 'accuracy.pkl')
createPickle(prec, 'prec.pkl')
createPickle(results, 'results.pkl')
createPickle(matrix, 'matrix.pkl')