In [3]:

    

import sys
import os
sys.path.append("/Users/ajeetjha/sandbox/sherlock/blackops/scripts")
import lib.dbUtil as dbUtil
import lib.genericUtil as gUtil
import lib.mongoUtil as mUtil
import pandas as pd
import numpy as np
from datetime import datetime, date, time, timedelta
import pprint


    

In [4]:

    

df = gUtil.readCsvToPD('d1_w1_data/first5_day_2017_2_16_d1.csv')


    

In [5]:

    

df.head()


    

    
Out[5]:






  

    

      

      
Unnamed: 0
      
pid
      
chaal_by_blind
      
avg_win_by_boot
      
avg_loss_by_boot
      
times_loaded
    
  
  

    

      
0
      
0
      
1006585138905
      
3.0
      
1
      
11
      
1
    
    

      
1
      
1
      
100642117074
      
0.0
      
0
      
0
      
0
    
    

      
2
      
2
      
100488274105
      
0.0
      
0
      
0
      
0
    
    

      
3
      
3
      
1005883985003
      
0.0
      
0
      
0
      
1
    
    

      
4
      
4
      
100624859958
      
1.0
      
0
      
4
      
1
    
  

In [4]:

    

def normaliseDataFrame(dataFrame):
    from sklearn import preprocessing
    min_max_scaler = preprocessing.MinMaxScaler()
    return pd.DataFrame(min_max_scaler.fit_transform(dataFrame))


    

In [5]:

    

#df1 = df[['chaals','wins','blinds','max_chips','times_loaded']]
df2 = df[['chaal_by_blind','avg_win_by_boot','avg_loss_by_boot','times_loaded']]


    

In [6]:

    

df3.head()


    

    




---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-6-334cab173724> in <module>()
----> 1 df3.head()

NameError: name 'df3' is not defined

In [6]:

    

df2.corr()


    

    
Out[6]:






  

    

      

      
chaal_by_blind
      
avg_win_by_boot
      
avg_loss_by_boot
      
times_loaded
    
  
  

    

      
chaal_by_blind
      
1.000000
      
0.290618
      
0.021134
      
0.120445
    
    

      
avg_win_by_boot
      
0.290618
      
1.000000
      
0.359030
      
0.198996
    
    

      
avg_loss_by_boot
      
0.021134
      
0.359030
      
1.000000
      
0.143982
    
    

      
times_loaded
      
0.120445
      
0.198996
      
0.143982
      
1.000000
    
  

In [7]:

    

%matplotlib inline


    

In [8]:

    

import matplotlib.pyplot as plt


    

In [9]:

    

plt.scatter(df1['wins'], df1['max_chips'], c= df1['times_loaded'] )


    

    




---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-9-56d4632c9a5c> in <module>()
----> 1 plt.scatter(df1['wins'], df1['max_chips'], c= df1['times_loaded'] )

NameError: name 'df1' is not defined

In [6]:

    

import numpy as np
from sklearn.cross_validation import train_test_split


    

In [7]:

    

y = df2.pop('times_loaded')
#df3 =  normaliseDataFrame(df2) # to normalize the data
X = df2.as_matrix()

X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.2)


    

In [8]:

    

# from sklearn import svm
# clf = svm.SVC() # 0.666021227768 accuracy on test set
# # Features: 'chaals','wins','blinds','max_chips' Labels:'times_loaded'
# clf.fit(X_train, y_train)


    

In [9]:

    

from sklearn import svm
clf = svm.SVC() 
# 0.703098106713 accuracy on test set, all install data, D1 return
# 0.708183010823 accuracy on test data, gplay install data, D1 return
# 0.657415375789 accuracy on test data, all install data, W1 Return
# Features: 'chaals','wins','blinds' Labels:'times_loaded'
clf.fit(X_train, y_train)


    

    
Out[9]:





SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,
  decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',
  max_iter=-1, probability=False, random_state=None, shrinking=True,
  tol=0.001, verbose=False)

In [13]:

    

# from sklearn import svm
# clf = svm.SVC() # 0.69714572576 accuracy on test set
# # Features: 'chaals','blinds' Labels:'times_loaded'
# clf.fit(X_train, y_train)


    

In [9]:

    

# from sklearn.naive_bayes import GaussianNB
# clf = GaussianNB()
# # Features: 'chaals','blinds', 'wins' Labels:'times_loaded' , accuracy 0.65
# clf.fit(X_train, y_train)


    

    
Out[9]:





GaussianNB()

In [9]:

    

# from sklearn import neighbors
# clf = neighbors.KNeighborsClassifier(15, weights='distance')
# # Features: 'chaals','blinds','wins' Labels:'times_loaded' , accuracy 0.68
# clf.fit(X_train, y_train)


    

    
Out[9]:





KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',
           metric_params=None, n_jobs=1, n_neighbors=15, p=2,
           weights='distance')

In [ ]:

    

# from sklearn import tree
# clf = tree.DecisionTreeClassifier()
# # Features: 'chaals','blinds','wins' Labels:'times_loaded' , accuracy 0.649129505192
# clf.fit(X_train, y_train)


    

In [19]:

    

# # Trying boosting
# from sklearn import svm
# from sklearn.ensemble import AdaBoostClassifier
# from sklearn.tree import DecisionTreeClassifier
# #dt = svm.SVC()
# dt = DecisionTreeClassifier()
# clf = AdaBoostClassifier(n_estimators=10, base_estimator=dt,learning_rate=1)
# # 0.703098106713 accuracy on test set, first5 games  data, D1 return
# # Features: 'chaals','wins','blinds' Labels:'times_loaded'
# clf.fit(X_train, y_train)


    

    
Out[19]:





AdaBoostClassifier(algorithm='SAMME.R',
          base_estimator=DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,
            max_features=None, max_leaf_nodes=None, min_samples_leaf=1,
            min_samples_split=2, min_weight_fraction_leaf=0.0,
            presort=False, random_state=None, splitter='best'),
          learning_rate=1, n_estimators=10, random_state=None)

In [10]:

    

y_pred = clf.predict(X_test)


    

In [11]:

    

from sklearn.metrics import accuracy_score
accuracy = accuracy_score(y_test, y_pred)


    

In [12]:

    

print(accuracy)


    

    




0.617864780392

In [13]:

    

from sklearn.metrics import confusion_matrix
confusion_matrix(y_test, y_pred)


    

    
Out[13]:





array([[4878, 3702],
       [2470, 4935]])

In [14]:

    

sum(y_test == 0)


    

    
Out[14]:





8580

In [17]:

    

df_validate = gUtil.readCsvToPD('d1_w1_data/first5_day_2017_1_27_d1.csv')


    

In [18]:

    

df_validate2 = df_validate[['chaal_by_blind','avg_win_by_boot','avg_loss_by_boot','times_loaded']]
y_val = df_validate2.pop('times_loaded')
X_val = df_validate2#.as_matrix()


    

In [19]:

    

y_val_pred = clf.predict(X_val)


    

In [20]:

    

from sklearn.metrics import accuracy_score
accuracy = accuracy_score(y_val, y_val_pred)
print(accuracy)


    

    




0.609063950651

In [36]:

    

len(y_val)


    

    
Out[36]:





77813

In [37]:

    

37534/77813


    

    
Out[37]:





0.48236155912251166

In [17]:

    

def writeClassifierToFile(clf, filename):
    from sklearn.externals import joblib
    pickle_path = os.path.join(gUtil.DPATH,filename)
    joblib.dump(clf, pickle_path)

def readClassifierFromFile(filename):
    from sklearn.externals import joblib
    pickle_path = os.path.join(gUtil.DPATH,filename)
    clf = joblib.load(pickle_path)
    return clf


    

In [18]:

    

writeClassifierToFile(clf,'d1_w1_data/W1_classifier.pkl')


    

In [19]:

    

clf2 = readClassifierFromFile('d1_w1_data/W1_classifier.pkl')


    

In [20]:

    

y_val_pred2 = clf.predict(X_val)


    

In [21]:

    

from sklearn.metrics import accuracy_score
accuracy = accuracy_score(y_val, y_val_pred2)
print(accuracy)


    

    




0.770287660371

In [23]:

    

sum(y_val_pred2 == 0)


    

    
Out[23]:





53629

In [46]:

    

36194/77813


    

    
Out[46]:





0.46514078624394384

In [ ]: