About the project
This was a project I did while I was getting started with Machine Learning. Coding Blocks had a competition where we had to classify Pokémons based on their names. It’s a multi-class classification problem and the Pokémons need to be classified as either Pikachu, Bulbasaur, or Charmander.
I learned how to implement the naïve KNN algorithm from scratch and use the Numpy for image processing. The biggest challenge was getting used to working with images as I had never done that before.
Some images from each category in the dataset
These are a few pictures from each category in the dataset provided. I scored an accuracy of over 97% in the final submission.
The github repository for the same can be found here.