In this thread, i'm building 3 Classifiers from scratch, for Digit-Recognition:
- Euclidean-Distance
- Gaussian-Naive-Bayes
- MLP-Classifier
I'm comparing the cross-val scores and plot their desicion surfaces, as well as the learning curves. in additon, for statistical reasons, i'm applying fine-tunning on Voting and Bagging models, to compare with mine.
In Version 2, i am going to present the mathematical base behind these implementations and observations.
In src directory, run:
python main.py