This is a part of Assignment 1 for Course DD2437 - Artificial Neural Networks and Deep Architectures at KTH.
This codebase use poetry as the python environment management tool,
please have it properly installed (with Python 3.8) before continuing.
Our dataset is a Mackey Glass series
There are 3 set of experiments:
- Grid search to find an optimal three-layer perceptron architecture
- Greedy search to fina an optimal network with best performance on noisy data
- Greedy search to find a good regularization
lambdaunder the present of additive Gaussian noise in the training data
More details could be found in experiment.py