def main():
from predict import get_test_data
X,y,test_X,test_Y =get_test_data()
print(X)
knn = KNNClassifier(n_neighbors=50,weight=distanceWeight)
knn.test(X,y,test_X,test_Y)
def get_proba(self,X,y,test_X,prefix=''):
if isfile(prefix+'bag_of_words_random_forest_train.pkl'):
rf_proba_train = pd.read_pickle(prefix+'bag_of_words_random_forest_train.pkl')
rf_proba_test = pd.read_pickle(prefix+'bag_of_words_random_forest_test.pkl')
else:
self.fit(X,y)
X = self._transform_data(X)
test_X=self._transform_data(test_X)
rf_proba_train = pd.DataFrame(self.random_forest_pipe.predict_proba(X))
rf_proba_train.to_pickle(prefix+'bag_of_words_random_forest_train.pkl')
rf_proba_test = pd.DataFrame(self.random_forest_pipe.predict_proba(test_X))
rf_proba_test.to_pickle(prefix+'bag_of_words_random_forest_test.pkl')
return [(rf_proba_train,rf_proba_test)]
def test(self,X,y,test_X,test_Y):
print("_"*80)
print(self.random_forest_pipe)
self.fit(X,y)
pred = self.predict(test_X)
print(metrics.accuracy_score(test_Y,pred))
rf = self.random_forest_pipe.named_steps['model']
print(rf.feature_importances_)
if __name__ == '__main__':
from predict import get_test_data
X,y,test_X,test_Y =get_test_data()
bow = BagOfWordsClassifier()
bow.test(X,y,test_X,test_Y)
