我正在尝试通过以下中型文章重新创建代码：https://medium.com/@rajatgupta310198/getting-started-with-neural-network-for-regression-and-tensorflow-58ad3bd75223
我正在努力复制结果，但遇到很多错误。
我对原始文章进行了一些调整：
scaler = MinMaxScaler（）＃为了规范化数据集，我摆脱了这一点。 删除了非规范化功能。
X_train = scaler.fit_transform（df_train.drop（['Close']，axis = 1）.as_matrix（）） y_train = scaler.fit_transform（df_train ['Close']。as_matrix（））
更改为
X_train = df_train.drop（['Close']，axis = 1） y_train = df_train ['关闭']
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import tensorflow as tf
df = pd.read_csv('data.csv') # read data set using pandas
print(df.info()) # Overview of dataset
df = df.drop(['Date'],axis=1) # Drop Date feature
df = df.dropna(inplace=False) # Remove all nan entries.
df = df.drop(['Adj Close','Volume'],axis=1) # Drop Adj close and volume feature
df_train = df[:1059] # 60% training data and 40% testing data
df_test = df[1059:]
X_train = df_train.drop(['Close'],axis=1)
y_train = df_train['Close']
X_test = df_test.drop(['Close'],axis=1)
y_test = df_test['Close']
def neural_net_model(X_data,input_dim):
W_1 = tf.Variable(tf.random_uniform([input_dim,10]))
b_1 = tf.Variable(tf.zeros([10]))
layer_1 = tf.add(tf.matmul(X_data,W_1), b_1)
layer_1 = tf.nn.relu(layer_1)
# layer 1 multiplying and adding bias then activation function
W_2 = tf.Variable(tf.random_uniform([10,10]))
b_2 = tf.Variable(tf.zeros([10]))
layer_2 = tf.add(tf.matmul(layer_1,W_2), b_2)
layer_2 = tf.nn.relu(layer_2)
# layer 2 multiplying and adding bias then activation function
W_O = tf.Variable(tf.random_uniform([10,1]))
b_O = tf.Variable(tf.zeros([1]))
output = tf.add(tf.matmul(layer_2,W_O), b_O)
# O/p layer multiplying and adding bias then activation function
# notice output layer has one node only since performing #regression
return output
xs = tf.placeholder("float")
ys = tf.placeholder("float")
output = neural_net_model(xs,3)
cost = tf.reduce_mean(tf.square(output-ys))
train = tf.train.GradientDescentOptimizer(0.001).minimize(cost)
with tf.Session() as sess:
# Initiate session and initialize all vaiables
sess.run(tf.global_variables_initializer())
saver = tf.train.Saver()
#saver.restore(sess,'yahoo_dataset.ckpt')
for i in range(100):
for j in range(X_train.shape[0]):
sess.run([cost,train],feed_dict= {xs:X_train[j,:].reshape(1,3), ys:y_train[j]})
# Run cost and train with each sample
c_t.append(sess.run(cost, feed_dict={xs:X_train,ys:y_train}))
c_test.append(sess.run(cost, feed_dict={xs:X_test,ys:y_test}))
print('Epoch :',i,'Cost :',c_t[i])
pred = sess.run(output, feed_dict={xs:X_test})
# predict output of test data after training
print('Cost :',sess.run(cost, feed_dict={xs:X_test,ys:y_test}))
y_test = denormalize(df_test,y_test)
pred = denormalize(df_test,pred)
#Denormalize data
plt.plot(range(y_test.shape[0]),y_test,label="Original Data")
plt.plot(range(y_test.shape[0]),pred,label="Predicted Data")
plt.legend(loc='best')
plt.ylabel('Stock Value')
plt.xlabel('Days')
plt.title('Stock Market Nifty')
plt.show()
if input('Save model ? [Y/N]') == 'Y':
saver.save(sess,'yahoo_dataset.ckpt')
print('Model Saved')
TypeError跟踪（最近一次通话） 在（）中 i在范围（100）中为6： 对于范围（X_train.shape [0]）中的j为7： ----> 8 sess.run（[cost，train]，feed_dict = {xs：X_train [j，：]。reshape（1,3），ys：y_train [j]}） 9＃运行成本并对每个样本进行培训 10 c_t.append（sess.run（cost，feed_dict = {xs：X_train，ys：y_train}））
〜\ AppData \ Local \ Continuum \ anaconda3 \ lib \ site-packages \ pandas \ core \ frame.py在 getitem 中（自身，密钥） 2683返回self._getitem_multilevel（键） 2684其他： -> 2685返回self._getitem_column（key） 2686 2687 def _getitem_column（self，key）：
_getitem_column中的〜\ AppData \ Local \ Continuum \ anaconda3 \ lib \ site-packages \ pandas \ core \ frame.py（self，key） 2690＃获取列 第2691章（许我倾城） -> 2692返回self._get_item_cache（key） 2693 2694＃重复的列并可能降低维数
_get_item_cache中的〜\ AppData \ Local \ Continuum \ anaconda3 \ lib \ site-packages \ pandas \ core \ generic.py（self，item） 2482“”“返回缓存的项目，该项目表示标签索引器。”“” 2483缓存= self._item_cache -> 2484 res = cache.get（项目） 2485如果res为None： 2486个值= self._data.get（item）
TypeError：不可散列的类型：'slice'