Hi Roberto,
When using neural networks for time series analysis in MATLAB, especially with the Deep Learning Toolbox, it's important to format your data appropriately to ensure effective training and prediction. Here's a guideline on the correct data format:
- For the input data, your 5000x1 array should be structured into sequences if you are using networks like LSTMs or RNNs. Each sequence serves as an input sample.
- The data should be in a format compatible with MATLAB's neural network functions. For LSTM networks, this often means using cell arrays where each cell contains a sequence of data points.
- Normalize your data to improve network performance. This involves scaling your data to a specific range, often [0, 1].
You can refer to the resource below for more information:
