-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathml-yelp.py
executable file
·144 lines (116 loc) · 4.65 KB
/
ml-yelp.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
#!/usr/bin/env python
import os, sys
import pandas as pd
import numpy as np
from sklearn import preprocessing, svm, cross_validation, tree, metrics
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.linear_model import SGDClassifier
from sklearn.model_selection import cross_val_score
from sklearn.naive_bayes import GaussianNB
from sklearn.naive_bayes import MultinomialNB
from sklearn.feature_selection import SelectKBest
from sklearn.feature_selection import chi2
from sklearn.preprocessing import MinMaxScaler
import matplotlib.pylab as plt
# Random seed for models
np.random.seed(0)
# Global vars
categorical_cols = ['neighborhood', 'type', 'state', 'city', 'name']
names = ['neighborhood', 'type', 'state', 'city', 'name',
'stars', 'postal_code', 'latitude', 'longitude',
'review_count', 'is_open']
def preprocess_data(df, cols):
processed_df = df.copy()
le = preprocessing.LabelEncoder()
processed_df[cols] = df[cols].apply(le.fit_transform)
return processed_df
def custom_processing(df):
# drop non numeric zip codes
filtered_df = df[df.postal_code.apply(lambda x: x.isnumeric())]
named_df = filtered_df.filter(names, axis=1)
named_df[categorical_cols] = named_df[categorical_cols].apply(lambda x: x.astype('category'))
named_df[['postal_code']] = named_df[['postal_code']].apply(lambda x: x.astype('int64'))
return named_df
def load_data():
cwd = os.getcwd()
return pd.read_json(path_or_buf=cwd+"/data/valid_json.json")
def plot_scores(data_dict):
lists = sorted(data_dict.items())
x_temp, y = zip(*lists)
x = np.arange(len(x_temp))
#print(x)
#print(y)
plt.bar(x, y, align='center', alpha=0.5)
plt.xticks(x, x_temp, rotation=45)
#plt.gcf().subplots_adjust(bottom=0.15)
plt.ylim(0.8520,0.8550)
plt.tight_layout()
plt.show()
def main():
raw_df = load_data()
df = custom_processing(raw_df)
df = preprocess_data(df, categorical_cols)
X = df.drop(['is_open'], axis=1).values
y = df['is_open'].values
# Scale data
scaler = MinMaxScaler()
X_scaled = scaler.fit_transform(X)
y_scaled = scaler.fit_transform(y)
# Feature Selection
# - Change k to number of k best features you want to keep
chi2k = SelectKBest(chi2, k=2)
X_new_train = chi2k.fit_transform(X_scaled, y_scaled)
mask = chi2k.get_support()
new_features = []
for bool, feature in zip(mask, names):
if bool:
new_features.append(feature)
# Create new dataframe from select features
# and scale features again.
X_pre_split = scaler.fit_transform(df[new_features].values)
y_pre_split = scaler.fit_transform(df['is_open'].values)
# Split data
X_train, X_test, y_train, y_test = cross_validation.train_test_split(X_pre_split,
y_pre_split,test_size=0.3)
# Plot data dict
plot_data = {}
# Decision Tree
clf_dt = tree.DecisionTreeClassifier(max_depth=10)
clf_dt.fit(X_train, y_train)
scores = cross_val_score(clf_dt, X_train, y_train)
print("Decision Tree: " + str(clf_dt.score(X_test ,y_test)))
print("Cross Validation Mean: " + str(scores.mean()) + "\n")
plot_data["Decision Tree"] = scores.mean()
# Random Forest
clf_rf = RandomForestClassifier(n_estimators=50)
clf_rf.fit(X_train, y_train)
scores = cross_val_score(clf_rf, X_train, y_train)
print("Random Forest: " + str(clf_rf.score(X_test,y_test)))
print("Cross Validation Mean: " + str(scores.mean()) + "\n")
plot_data["Random Forest"] = scores.mean()
# Gradient Boosting Classifier
clf_gb = GradientBoostingClassifier(n_estimators=200, learning_rate=0.1)
clf_gb.fit(X_train, y_train)
scores = cross_val_score(clf_gb, X_train, y_train)
print("Gradient Boosting: " + str(clf_gb.score(X_test,y_test)))
print("Cross Validation Mean: " + str(scores.mean()) + "\n")
plot_data["GBoost"] = scores.mean()
# Naive Bayes Classifiers
# Gaussian NB
clf_gnb = GaussianNB()
clf_gnb.fit(X_train, y_train)
scores = cross_val_score(clf_gnb, X_train, y_train)
print("Gaussian NB: " + str(clf_gnb.score(X_test,y_test)))
print("Cross Validation Mean: " + str(scores.mean()) + "\n")
plot_data["Gaussian NB"] = scores.mean()
# Multinomial NB
clf_mnb = MultinomialNB()
clf_mnb.fit(X_train, y_train)
scores = cross_val_score(clf_mnb, X_train, y_train)
print("Multinomial NB: " + str(clf_mnb.score(X_test,y_test)))
print("Cross Validation Mean: " + str(scores.mean()) + "\n")
plot_data["Multinomial NB"] = scores.mean()
plot_scores(plot_data)
if __name__ == "__main__":
main()