A simple implementation of artificial neural networks using numpy as randomizers and python function only. This could still be improved by using faster functions in Numpy. Additionally, Synthethic Minority Oversampling Technique (SMOTE) was also used to handle class imbalance.
Data used was removed but a report is provided for visual appreciation.
Performance of varying polynomial degrees
- numpy
- collections
- matplotlib
- imblearn
- timeit
- sklearn