notebook.community

Edit and run 





In [3]:



    


import pandas as pd
import numpy as np
import pydotplus
import matplotlib.pyplot as plt
%matplotlib inline

from sklearn import preprocessing
from sklearn.tree import DecisionTreeRegressor
from sklearn.metrics import mean_squared_error
from sklearn.tree import export_graphviz
from sklearn.model_selection import train_test_split
from IPython.display import Image

import data





In [4]:



    


data = pd.read_csv('../../Data/WeatherOutagesAllJerry.csv')



    











In [5]:



    


data = data.dropna(how = 'all')



    











In [6]:



    


data.head(5)



    


















    
Out[6]:










  
    

      
      Date
      Total_outages
      Equipment
      Trees
      Animals
      Lightning
      Day_length_hr
      Max_temp_F
      Avg_Temp_F
      Min_temp_F
      ...
      Avg_windspeed_mph
      Max_windgust_mph
      Precipitation_in
      Events
      Event_fog
      Event_rain
      Event_snow
      Event_thunderstorm
      Event_Hail
      Event_Tornado
    

  
  
    

      0
      9/11/00
      0.0
      0.0
      0.0
      0.0
      0.0
      12.783333
      66.0
      58.0
      50.0
      ...
      2.0
      9.0
      0.01
      Fog
      1.0
      0.0
      0.0
      0.0
      0.0
      0.0
    

    

      1
      9/12/00
      2.0
      1.0
      0.0
      1.0
      0.0
      12.716667
      75.0
      62.0
      52.0
      ...
      4.0
      9.0
      0.00
      NaN
      0.0
      0.0
      0.0
      0.0
      0.0
      0.0
    

    

      2
      9/13/00
      1.0
      1.0
      0.0
      0.0
      0.0
      12.666667
      77.0
      64.0
      54.0
      ...
      7.0
      25.0
      0.00
      Fog
      1.0
      0.0
      0.0
      0.0
      0.0
      0.0
    

    

      3
      9/14/00
      0.0
      0.0
      0.0
      0.0
      0.0
      12.616667
      84.0
      71.0
      60.0
      ...
      12.0
      9.0
      0.00
      NaN
      0.0
      0.0
      0.0
      0.0
      0.0
      0.0
    

    

      4
      9/15/00
      1.0
      1.0
      0.0
      0.0
      0.0
      12.550000
      73.0
      66.0
      59.0
      ...
      5.0
      9.0
      0.00
      NaN
      0.0
      0.0
      0.0
      0.0
      0.0
      0.0
    

  



5 rows × 27 columns



















In [7]:



    


np.shape(data)



    


















    
Out[7]:








(5664, 27)

















In [8]:



    


#data = data.fillna(0)
#data = data.round(4)



    











In [9]:



    


train,test=train_test_split(data,test_size=0.1,random_state=567)



    











In [17]:



    


x_train = train.iloc[:,6:19].join(train.iloc[:,21:])
y_train = train.iloc[:,1]

x_test = test.iloc[:,6:19].join(test.iloc[:,21:])
y_test = test.iloc[:,1]



    











In [18]:



    


# data normalization
x_train = preprocessing.normalize(x_train)
#y_train = preprocessing.normalize(y_train)

x_test = preprocessing.normalize(x_test)
#y_test = preprocessing.normalize(y_test)



    











In [ ]:



    


# data standardization
x_train_std = preprocessing.scale(x_train)
y_train_std = preprocessing.scale(y_train)



    











In [19]:



    


x_train.size



    


















    
Out[19]:








96843

















In [20]:



    


y_test.size



    


















    
Out[20]:








567

Decision Tree





In [40]:



    


predictor = DecisionTreeRegressor(criterion='mse',max_leaf_nodes=6)



    











In [41]:



    


clf = predictor.fit(x_train,y_train)



    











In [42]:



    


y_trainpred = predictor.predict(x_train)
y_pred = predictor.predict(x_test)



    











In [43]:



    


score = mean_squared_error(y_test, y_pred)
print("Train error",mean_squared_error(y_train,y_trainpred))
print("Test error",mean_squared_error(y_test,y_pred))



    


















    






Train error 3.36246724318
Test error 3.71908873168

















In [44]:



    


tree_dot = export_graphviz(clf,out_file = 'tree_all_normalized.dot')



    











In [45]:



    


graph = pydotplus.graphviz.graph_from_dot_file('tree_all_normalized.dot')
Image(graph.create_png())



    


















    
Out[45]:
























In [46]:



    


plt.figure(figsize=(5,5))
plt.scatter(y_test,y_pred,alpha=0.2)
plt.xlabel('real test value')
plt.ylabel('predicted test value')
plt.plot(np.arange(-5,20),np.arange(-5,20))



    


















    
Out[46]:








[<matplotlib.lines.Line2D at 0x10fe74cf8>]

Random Forests





In [53]:



    


from sklearn.ensemble import RandomForestRegressor, ExtraTreesRegressor
from sklearn.cross_validation import ShuffleSplit
from sklearn.learning_curve import validation_curve



    











In [55]:



    


rf = RandomForestRegressor(n_estimators=100)
rf.fit(x_train,y_train)



    


















    
Out[55]:








RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=None,
           max_features='auto', max_leaf_nodes=None,
           min_impurity_split=1e-07, min_samples_leaf=1,
           min_samples_split=2, min_weight_fraction_leaf=0.0,
           n_estimators=100, n_jobs=1, oob_score=False, random_state=None,
           verbose=0, warm_start=False)

















In [56]:



    


y_trainpred = rf.predict(x_train)
y_pred = rf.predict(x_test)



    











In [57]:



    


print("Train error",mean_squared_error(y_train,y_trainpred))
print("Test error",mean_squared_error(y_test,y_pred))



    


















    






Train error 0.587340357073
Test error 2.99659929453

















In [62]:



    


plt.figure(figsize=(5,5))
plt.scatter(y_test,y_pred,alpha=0.2)
plt.xlabel('real test value')
plt.ylabel('predicted test value')
plt.plot(np.arange(0,20),np.arange(0,20))



    


















    
Out[62]:








[<matplotlib.lines.Line2D at 0x116be0fd0>]

Extra Trees





In [59]:



    


from sklearn.ensemble import (RandomForestRegressor, ExtraTreesRegressor,
                              AdaBoostRegressor)



    











In [72]:



    


et = ExtraTreesRegressor(n_estimators=100)
et.fit(x_train,y_train)



    


















    
Out[72]:








ExtraTreesRegressor(bootstrap=False, criterion='mse', max_depth=None,
          max_features='auto', max_leaf_nodes=None,
          min_impurity_split=1e-07, min_samples_leaf=1,
          min_samples_split=2, min_weight_fraction_leaf=0.0,
          n_estimators=100, n_jobs=1, oob_score=False, random_state=None,
          verbose=0, warm_start=False)

















In [73]:



    


y_trainpred = et.predict(x_train)
y_pred = et.predict(x_test)



    











In [74]:



    


print("Train error",mean_squared_error(y_train,y_trainpred))
print("Test error",mean_squared_error(y_test,y_pred))



    


















    






Train error 0.0
Test error 2.76110740741

















In [75]:



    


plt.figure(figsize=(5,5))
plt.scatter(y_test,y_pred,alpha=0.2)
plt.xlabel('real test value')
plt.ylabel('predicted test value')
plt.plot(np.arange(0,20),np.arange(0,20))



    


















    
Out[75]:








[<matplotlib.lines.Line2D at 0x1169c4b70>]

AdaBoost





In [76]:



    


ab = AdaBoostRegressor()
ab.fit(x_train,y_train)



    


















    
Out[76]:








AdaBoostRegressor(base_estimator=None, learning_rate=1.0, loss='linear',
         n_estimators=50, random_state=None)

















In [77]:



    


y_trainpred = ab.predict(x_train)
y_pred = ab.predict(x_test)



    











In [78]:



    


print("Train error",mean_squared_error(y_train,y_trainpred))
print("Test error",mean_squared_error(y_test,y_pred))



    


















    






Train error 10.8160484877
Test error 11.9008246196

















In [79]:



    


plt.figure(figsize=(5,5))
plt.scatter(y_test,y_pred,alpha=0.2)
plt.xlabel('real test value')
plt.ylabel('predicted test value')
plt.plot(np.arange(0,20),np.arange(0,20))



    


















    
Out[79]:








[<matplotlib.lines.Line2D at 0x1199c1978>]










    
























In [ ]:

Content source: rkastilani/PowerOutagePredictor

Similar notebooks:

notebook.community | gallery | about