tf.Graph.as_graph_def(sess.graph))
for epoch in range(PARAM_NUM_EPOCH):
#print "Training in epoch: ", epoch
#log_file_object.write("Training in epoch: "+str(epoch)+"\r\n")
#for start, end in zip(range(0, len(trX), 128), range(128, len(trX), 128)):
sess.run(train_op, feed_dict={X: trX, Y: trY})
cost_log = sess.run(cost, feed_dict={X: trX, Y: trY})
log_file_object.write(str(cost_log) + "\r\n")
log_file_object.flush()
predict_result = sess.run(predict_op, feed_dict={X: teX, Y: teY})
statistic_file_object.write("Training in epoch: " + str(epoch) +
"\r\n")
statistic_file_object.write("R2 : \r\n")
R2 = data_loader.R2(np.array(predict_result), teY)
statistic_file_object.write(str(R2) + "\r\n")
statistic_file_object.flush()
# predict option
#batch_xs, batch_ys = drag_data.test.next_batch(128)
predict_result_temp = sess.run(predict_op, feed_dict={X: teX, Y: teY})
#print predict_result
predict_result = []
for pred in predict_result_temp:
predict_result.append(pred[0])
# write file
statistic_file_object.write("\r\nFinal : \r\n")
statistic_file_object.write("True Activities: \r\n")
import numpy as np
import cPickle as pickle
import data_loader
data_file_train = "/Users/peter/Documents/Work/data/drag_design/NK1_training_disguised.csv"
data_file_test = "/Users/peter/Documents/Work/data/drag_design/NK1_test_disguised.csv"
drag_data = data_loader.read_data_sets(data_file_train, data_file_test, 1000)
trX, trY, teX, teY = drag_data.train.descriptors, drag_data.train.activities, drag_data.test.descriptors, drag_data.test.activities
# Random Forest Classifier
def random_forest_regressor(train_x, train_y):
from sklearn.ensemble import RandomForestRegressor
model = RandomForestRegressor(n_estimators=200)
model.fit(train_x, train_y)
return model
if __name__ == '__main__':
num_train, num_feat = trX.shape
num_test, num_feat = teX.shape
model = random_forest_regressor(trX, trY)
predict = model.predict(teX)
#accuracy = metrics.accuracy_score(teY, predict)
R2 = data_loader.R2(np.array(predict), teY)
print "R : ", R2