Trying to copy a LSTM model - it's not working

Question

I'm trying to copy a LSTM model that I found from here: Stock Market-Predict volume with LSTM model

I'm getting stuck on the last line of code. Specifically, this is what it tells me:

I literally know next to nothing about code just basic python.

If you want to, I can add you to like the google drive collaboratory so you could look at the whole code.

Answer 1

The arrays aren't the same size, which is required for the pyplot.plot function to work. So just cut less items out of the data array like so from:

plt.plot(data.index[-640:], test_y, color='blue',label='Actual')
plt.plot(data.index[-640:], yPredict, alpha=0.7, color='red',label='Predict')

to

plt.plot(data.index[-300:], test_y[-300:], color='blue',label='Actual')
plt.plot(data.index[-300:], yPredict[-300:], alpha=0.7, color='red',label='Predict')

just to explain this even more what its saying is that the size of the arrays aren't the same, so what I did was cut out the items equally + less because it said that test_y had 372 items, however I don't know if this is the same for yPredict, it could be even smaller. good luck.

Answer 2

To add to the other answer, the train and test sets are built here:

num = int(data_reframed.shape[0]*0.8)
value = data_reframed.values
train = value[:num, :]
test = value[num:, :]

train_x, train_y = train[:, :-1], train[:, -1]
test_x, test_y = test[:, :-1], test[:, -1]

While yPredict is calculated here:

yPredict = Model.predict(test_x)

Hence, the shape of yPredict and test_y has to be the same, and that has to be (data_reframed.shape[0] - num,), since num is the size of the training set. Then, due to series_to_supervised?, datahas one row more thandata_reframed`. So, you can write:

plt.plot(data.index[num+1:], test_y, color='blue',label='Actual')
plt.plot(data.index[num+1:], yPredict, alpha=0.7, color='red',label='Predict')

And plot everything.