Many to One LSTM Keras Python

Question

I have the below dataset for a process of 5 features to produce 1 output and 244000 rows. The output is in column 7. Each input is sampled every 1/10 second. I wanted to predict the 7th column (value 1 (t+1)) with an LSTM network with a timestep of 5.

My question is: What should be my 3D input tensor parameters? Is [244000, 5, 5] correct? And how can I reshape my dataset to this shape?

While trying it with the reshape numpy function I get an error.

Answer 1

If I read correctly, at each timestep the model gets a 5 feature input and you take 5 such timesteps. That would make the first dimension 244000/5 = 48800