Machine Learning Explorer

Inspiration

The project was built with the intent to demonstrate the importance of good feature selection when writing machine learning models

What it does

How we built it

I used anaconda for virtual environment management, kaggle for the fruit images dataset that we used for training and testing, opencv for image processing, and sci-kit learn for model construction.

The piece of art in the image section was AI generated .

Challenges we ran into

It was very difficult to stay awake the whole night, but I managed it!

Accomplishments that we're proud of

The final model achieved a very high accuracy when all features were implemented.

What we learned

This was the first time I used a kmeans clustering algorithm to aid in classification.

What's next for Machine Learning Explorer

There needs to be a full implementation of the feature filter tool to allow the user to remove and add features to see how different features affect the final accuracy of the model. It would also be nice to improve the user interface and to allow the user to try different types of models on the extracted features. It would also be cool to add in some more datasets so the user can see how certain algorithms perform for different tasks like object detection, natural language processing, and pattern/structure extrapolation.

Built With

• dash
• kaggle
• python
• scikit-learn