This is an implementation of logistic regression (machine learning) in Typescript. The file includes a number of indirectly-related helper functions which can find the mean of and normalize a 2D matrix, among other things. I'm 80% certain it works (some parts of it don't have tests yet).
The main logistic() function takes a 2D matrix of inputs and a 1D vector of outputs (along with some other optional parameters, see the code for details). It returns a function that predicts the label (0 or 1) of a new observation (1D vector of same length as the # of columns the 2D matrix of training inputs has).
I made this primarily to reinforce my understanding of basic machine learning concepts, but if you want to use it to run logistic regression on the web, that's OK with me.
Included with the main.ts file is some code that performs logistic regression, using this API, on the highly-popular heart disease dataset. This serves as sort-of a proof-of-concept/benchmark, and was coded in about two minutes. Please don't hold its crudeness against the rest of the project