Try it out - project Link (use a laptop/PC)
This is an interactive learning and exploration tool that lets user pick an algorithm and a dataset of his choice and lets him/her play arround with the algorithm's hyperparameters while showing instantly how any change in hyperParameters values affect accuracy & decision boundry computed by the algorithm.
Often times when people first start out with ML getting an intuitive idea about how model's paraneter affect the output and accuracy can be vague. This tool can be used to understand how model's parameters afftect accuracy and various other things in ML like overfitting, underfitting, generalization and decision boundry. Not only that it can also be used as a tool to explore different sets of hyper parameters and how each affects the model's accuracy.
This project was made using python and following python libraries - streamLit scikit-learn numpy matplotlib
- How to create an app using streamLit and deploy it
- working with Sklearn
I've had a few extra idea that i'll add to this project in future, like -
- Using optuna to do automatic hyperParamter optimization
- option to let user upload his/her own data
- descriptions of ML algorithms, so that it can be a better learning resource
- add more hyperParameter options for tweaking the algorithm