How to transfrom back from the MinMaxScaler?

Question

When I use the lstm to predict stock, I totally succeed to predict something.But, When I use the inverse_transform to make the number of predictions become normal from MinMaxScaler, it shows me error.

import pandas as pd
import numpy as np
import datetime as datetime
import pandas_datareader as pdr

start = datetime.datetime(2015,6,29)
end = datetime.datetime(2018,6,29)
df_2330 = pdr.DataReader("2330.TW", 'yahoo', start=start,end=end)
data = df_2330.drop(['Adj Close'],axis=1)
change = data.Close.diff()
data['Change'] = change
change.fillna(change.mean(),inplace=True)
data['RD'] = data.Close.pct_change()*100
data = data.dropna(axis=0,how='any')
df = data.drop(['Change'],axis=1)
train = df[0:491]
test = df[491:]

from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
train = pd.DataFrame(scaler.fit_transform(train), columns=df.columns)
test = pd.DataFrame(scaler.fit_transform(test), columns=df.columns)

import tqdm
n = 30 
feature_names = list(train.drop('RD', axis=1).columns)
X = []
y = []
indexes = []
norm_data_x = train[feature_names]
for i in tqdm.tqdm_notebook(range(0,len(train)-n)): 
    X.append(norm_data_x.iloc[i:i+n]. values) 
    y.append(train['RD'].iloc[i+n-1]) 
    indexes.append(train.index[i+n-1]) 
X=np.array(X) 
y=np.array(y)

import keras
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import LSTM
from keras.layers import Dropout

n_steps = 30 
n_features = 5
model = Sequential()
model.add(LSTM(50,activation='relu', return_sequences=True, input_shape = (n_steps, n_features)))
model.add(Dropout(0.2))
model.add(LSTM(units = 50,activation='relu',return_sequences = True))
model.add(Dropout(0.2))
model.add(LSTM(units = 50, return_sequences = True))
model.add(Dropout(0.2))
model.add(LSTM(units = 50))
model.add(Dropout(0.2))
model.add(Dense(units = 1))
model.add(Dense(1))
model.compile(optimizer = 'adam', loss = 'mse' , metrics=['mse','mape'])
history = model.fit(X,y,batch_size=32,epochs=20)

import tqdm
n = 30
feature_names = list(test.drop('RD', axis=1).columns)
X = []
y = []
indexes = []
norm_data_x = test[feature_names]
for i in tqdm.tqdm_notebook(range(0,len(test)-n)): 
    X.append(norm_data_x.iloc[i:i+n].values) 
    y.append(test['RD'].iloc[i+n-1]) #現有資料+30天的Y
    indexes.append(test.index[i+n-1]) #Y的日期
X=np.array(X)
y=np.array(y)
predictions = model.predict(X)
predictions = pd.DataFrame(predictions).rename(columns={0: '預測值'})
Y_test = pd.DataFrame(y).rename(columns={0: '實際值'})
final = pd.concat([predictions,Y_test],axis=1)
final['mae'] = abs(final['預測值'] - final['實際值'])
print(final.columns)
norm_data = pd.DataFrame(scaler.inverse_transform(final), columns=final.columns, index=final.index)
print(final)

  File "C:\Users\apple\.spyder-py3\try\untitled5.py", line 105, in <module>
    norm_data = pd.DataFrame(scaler.inverse_transform(final), columns=final.columns, index=final.index)

  File "C:\ProgramData\Anaconda3\envs\tensorflowenv\lib\site-packages\sklearn\preprocessing\_data.py", line 529, in inverse_transform
    X -= self.min_

ValueError: operands could not be broadcast together with shapes (214,3) (6,) (214,3)