-
Notifications
You must be signed in to change notification settings - Fork 0
/
old.py
165 lines (118 loc) · 5.66 KB
/
old.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
import pickle
import pandas as pd
from sklearn import metrics
'''
Pre function contains the old code if someone wants to know how the independent (X) and depenent (y) features were created.
already_built() is used to denote the code for already built fun.
build_model() is used to built a new model without disturbing the old model
'''
def pre():
df = pd.read_csv("MyCall_Data_September_2019_cleaning.csv")
df.drop(['Latitude', 'Longitude'], axis='columns', inplace=True)
df.rename(columns={
'In Out Travelling': 'InOut', 'Network Type': 'NetworkType', 'Call Drop Category': 'CallDropCategory',
'State Name': 'StateName'
}, inplace=True)
obj = {
'Operator': {'RJio': 1, 'Airtel': 2, 'Idea': 3, 'Other': 4, 'Vodafone': 5, 'BSNL': 6, 'MTNL': 7},
'InOut': {'Indoor': 1, 'Outdoor': 2, 'Travelling': 3},
'NetworkType': {'4G': 1, '3G': 2, '2G': 3},
'CallDropCategory': {'Satisfactory': 1, 'Poor Voice Quality': 2, 'Call Dropped': 3},
'StateName': {}
}
count = 1
for x in df['StateName'].unique():
obj['StateName'][x] = count
count += 1
# Do not run this code two times, otherwise the map function will convert all non-matched values to NaN
df['Operator'] = df['Operator'].map(
{'RJio': 1, 'Airtel': 2, 'Idea': 3, 'Other': 4, 'Vodafone': 5, 'BSNL': 6, 'MTNL': 7})
df['InOut'] = df['InOut'].map({'Indoor': 1, 'Outdoor': 2, 'Travelling': 3})
df['NetworkType'] = df['NetworkType'].map({'4G': 1, '3G': 2, '2G': 3})
df['CallDropCategory'] = df['CallDropCategory'].map({'Satisfactory': 1, 'Poor Voice Quality': 2, 'Call Dropped': 3})
df['StateName'] = df['StateName'].map(obj['StateName'])
X = df.columns.tolist()
X.remove('CallDropCategory')
X = df[X]
X.to_pickle('Models/X.pkl')
y = df['CallDropCategory']
y.to_pickle('Models/y.pkl')
def already_built(X_train, X_test, y_train, y_test):
from sklearn.naive_bayes import GaussianNB
gaussNb = GaussianNB()
gaussNb.fit(X_train.values, y_train.values)
# predicting test set results
y_pred = gaussNb.predict(X_test)
# saving the model as pickel module
with open('Models/gaussNb_model.pkl', 'wb') as file:
pickle.dump(gaussNb, file)
print("\nGaussian Naive Bayes model accuracy(in %):", metrics.accuracy_score(y_test, y_pred) * 100)
print(metrics.classification_report(y_test, y_pred))
from sklearn.naive_bayes import MultinomialNB
MultiNb = MultinomialNB()
MultiNb.fit(X_train.values, y_train.values)
# saving the model as pickel module
with open('Models/multiNb_model.pkl', 'wb') as file:
pickle.dump(MultiNb, file)
y_pred = MultiNb.predict(X_test)
print("\Multinomail Naive Bayes model accuracy(in %):", metrics.accuracy_score(y_test, y_pred) * 100)
print(metrics.classification_report(y_test, y_pred))
from sklearn import tree
decisionTree = tree.DecisionTreeClassifier()
decisionTree.fit(X_train.values, y_train.values)
y_pred = decisionTree.predict(X_test)
with open('Models/decisionTree_model.pkl', 'wb') as file:
pickle.dump(decisionTree, file)
print("\nDecision Tree(in %):", metrics.accuracy_score(y_test, y_pred) * 100)
print(metrics.classification_report(y_test, y_pred))
from sklearn.ensemble import RandomForestClassifier
randomForest = RandomForestClassifier(n_estimators=25)
randomForest.fit(X_train.values, y_train.values)
y_pred = randomForest.predict(X_test)
with open('Models/randomForest_model.pkl', 'wb') as file:
pickle.dump(randomForest, file)
print("\nRandomForest(in %):", metrics.accuracy_score(y_test, y_pred) * 100)
print(metrics.classification_report(y_test, y_pred))
from sklearn.neighbors import KNeighborsClassifier
knn = KNeighborsClassifier(n_neighbors=6)
knn.fit(X_train.values, y_train.values)
y_pred = knn.predict(X_test)
with open('Models/knn_model.pkl', 'wb') as file:
pickle.dump(knn, file)
print("\nKNN (in %):", metrics.accuracy_score(y_test, y_pred) * 100)
print(metrics.classification_report(y_test, y_pred))
from sklearn.neural_network import MLPClassifier
MLP = MLPClassifier()
MLP.fit(X_train.values, y_train.values)
y_pred = MLP.predict(X_test)
with open('Models/MLP_model.pkl', 'wb') as file:
pickle.dump(MLP, file)
print("\nMLP (in %):", metrics.accuracy_score(y_test, y_pred) * 100)
print(metrics.classification_report(y_test, y_pred))
from sklearn.svm import SVC
svm = SVC(gamma=2)
svm.fit(X_train.values, y_train.values)
y_pred = svm.predict(X_test)
with open('Models/svm_model.pkl', 'wb') as file:
pickle.dump(svm, file)
print("\nSVM (in %):", metrics.accuracy_score(y_test, y_pred) * 100)
print(metrics.classification_report(y_test, y_pred))
def build_model(X_train, X_test, y_train, y_test):
from sklearn.naive_bayes import MultinomialNB
MultiNb = MultinomialNB()
MultiNb.fit(X_train.values, y_train.values)
# saving the model as pickel module
with open('Models/multiNb_model.pkl', 'wb') as file:
pickle.dump(MultiNb, file)
y_pred = MultiNb.predict(X_test)
print("\Multinomail Naive Bayes model accuracy(in %):", metrics.accuracy_score(y_test, y_pred) * 100)
print(metrics.classification_report(y_test, y_pred))
def main():
df = pd.read_pickle("Models/saved_df")
X = pd.read_pickle("Models/X.pkl")
y = pd.read_pickle("Models/y.pkl")
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=1)
build_model(X_train, X_test, y_train, y_test)
if __name__ == "__main__":
main()