def main(_):
for i in [0]:
for j in [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]:
# for j in [1, 2, 3, 4, 5]:
# for j in [6, 7, 8, 9, 10]:
print('hhhe_ebdt{}_index{}.txt'.format(i, j))
#sys.stdout = open('ebdt'+str(i)+'_index'+str(j)+'.txt', 'w')
# for each label
y_true = []
y_pred_p = []
y_pred = []
sum_time_cost = 0.
for k in range(9):
print('hhhe_ebdt{}_index{}_label{}.txt'.format(i, j, k))
g = tf.Graph()
with g.as_default():
obj = Bi_LSTM_Att_BC_model(
test_index=j,
embedding_type=i,
label_index=k,
scope='hhhe_ebdt{}_index{}_label{}.txt'.format(
i, j, k))
time_cost, cur_y_true, best_y_pred_p = obj.run()
sum_time_cost += time_cost
# print "cur_y_true[0] : "
print cur_y_true[0]
# print "best_y_pred_p[9] : "
print best_y_pred_p[0]
cur_y_true = np.argmax(cur_y_true, axis=1).reshape(-1, 1)
best_y_pred = np.argmax(best_y_pred_p, axis=1).reshape(-1, 1)
best_y_pred_p = best_y_pred_p[:, 1].reshape(-1, 1)
if k == 0:
y_true = cur_y_true
y_pred = best_y_pred
y_pred_p = best_y_pred_p
else:
y_true = np.concatenate((y_true, cur_y_true), axis=1)
y_pred = np.concatenate((y_pred, best_y_pred), axis=1)
y_pred_p = np.concatenate((y_pred_p, best_y_pred_p),
axis=1)
Emotion_eval(y_true, y_pred, y_pred_p)
Emo_eval(y_true, y_pred)
print y_true[0]
print y_pred[0]
print y_pred_p[0]
# np.savetxt('For_calibrated_B'+str(j)+'_share2.txt', y_pred, fmt="%d", delimiter=",")
print "Finish the {}th experiment!".format(j)
print "sum of the time cost for each binary classifier: {}".format(
sum_time_cost)
def new_test():
feed_dict = {
self.x: ev_x,
self.doc_len: ev_doc_len,
self.keep_prob: 1.0,
}
y_true = ev_y
y_pred_p = sess.run(prob, feed_dict=feed_dict)
y_pred = calibrated_label_ranking(y_pred_p, For_calibrated_B)
Emotion_eval(y_true, y_pred, y_pred_p)
def new_test():
feed_dict = {
self.x: te_x,
self.doc_len: te_doc_len,
self.keep_prob: 1.0,
}
y_true = te_y
y_pred_p = sess.run(prob, feed_dict=feed_dict)
# y_pred = calibrated_label_ranking(y_pred_p, For_calibrated_B)
y_pred = np.ceil(y_pred_p - 1.0 / 9)
Emotion_eval(y_true, y_pred, y_pred_p)
def new_test():
feed_dict = {
self.x: te_x,
self.doc_len: te_doc_len,
self.keep_prob: 1.0,
}
y_true = np.ceil(te_y)
y_pred_p = sess.run(prob, feed_dict=feed_dict)
#修改 ： 采用calibrated label ranking 算法
# y_pred = np.ceil(y_pred_p-1.0/8)
y_pred = calibrated_label_ranking(y_pred_p)
Emotion_eval(y_true, y_pred, y_pred_p)
def new_test():
feed_dict = {
self.x: te_x,
self.doc_len: te_doc_len,
self.keep_prob: 1.0,
}
y_true = te_y[:, 1 : : 2]
y_pred_p = sess.run(prob, feed_dict=feed_dict)[:, 1 : : 2]
y_pred = np.ceil(y_pred_p-1.0/2)
''' test '''
# print('y_true[0] : {}'.format(y_true[0]))
# print('y_pred_p[0] : {}'.format(y_pred_p[0]))
# print('y_pred[0] : {}'.format(y_pred[0]))
Emotion_eval(y_true, y_pred, y_pred_p)