Description:

To learn how to use the BP algorithm to train a multi-layer network. The provided code should be written by Python.

Requirement:

using the BP algorithm design a three layers network (i.e., three hidden layers) and carefully train the network with one classification dataset selected from the UCI dataset library. (https://archive.ics.uci.edu/ml/datasets.php). Provide your testing accuracy on the selected dataset.

Select a regression dataset from the UCI dataset library. Then using BP to train a three layers network and provide your testing/training root mean square error.

Please select the SGDM optimizer in the training process.

Any built-in function or toolboxes should not be used to complete the assignment.

Please attach dataset link or file with code.