I’ve seen other tutorials did this — splitting sequence into multiple sub-sequences for CNN..not

Original article was published on Deep Learning on Medium


…uential model. The input shape would be 24 time steps with 1 feature for a simple univariate model. Within the convolutional layer, I didn’t further split the sequence into multiple subsequences but keep the timesteps to be 24 instead. And I chose the kernel size to be 3. The dense layer has 24 neurons to produce 24 output number. B…