PolyReg is a simple web tool for doing Polynomial Regression analysis on a set of data points. It takes a CSV of x,y coordinates as training data and computes an N-degree polynomial using Ordinary Least Squares with L2 Regularization. It takes another CSV of data points as testing data, from which it computes the testing error against the selected polynomial.
This was written ages ago, sometime around 2012. You can tell how old it is from the choice of technology: it uses jQuery, jQuery-UI, and Flot for plotting. It is built on top of the codebase of my reCalc project, which is why they look so similar.
It used to require a Ruby backend which would do all of the computation, and return a polynomial. But thanks to wonderful new purely clientside matrix libraries like Math.js, I was able to port the backend to be entirely clientside. This makes it much simpler for me to host and keep up on this site.
Demo Link
It doesn't do anything fancy like Stochastic Gradient Descent for finding a best fit polynomial, it simply uses Ordinary Least Squares.
I'll try to expand this section a little bit sometime in the future.