Question

[cancerberoSgx]

Hello there, sorry to ask here, but could you briefly explain in which kind of problems this is used ? I'm a really beginner with ML, and I wonder if this is helpful for calculate classification kernel initial parameters (like documented https://www.csie.ntu.edu.tw/~cjlin/papers/guide/guide.pdf ? I'm trying to have this part clear before I proceed and I can't find any clear documentation or example on (https://github.com/mljs). When I do I promise to write some wiki or README regarding this. I see that examples don't detail on this and I see you are making calculating this like using data-set-iris like in https://github.com/mljs/pca/blob/master/src/__tests__/iris.js, https://github.com/mljs/knn/blob/master/src/__tests__/test.js. While is fine for testing, I don't think that's the recommended way to load datasets for users right?

Do you know if in this space project's there are tools like grid.py or easy.py as documented in https://www.csie.ntu.edu.tw/~cjlin/papers/guide/guide.pdf that do it automatically?. I guess no and is responsibility of each project to support that (or not). Am I right (i'm at that stage right now) ?

I would like to be sure regarding this since I read more that once that for classifiers-like that part is critical. Also because while I learn I'm planning to write some getting started notes
how the experience documented in previous PDF would be using projects here. And perhaps when I feel more secure some comment line tools which I think would simplify things a lot and would be easy to implement in node.js - even using chalk or blessed to have a killer CLI for training and testing.

Regarding this project I found a TypeScript implementation the only JS like mentioned in the original author's website (http://paulbourke.net/papers/conrec/conrec.ts) (probably you already knew) but just in case you are planning to write some typings or if the current one is older that that wanted to comment.

One last tip, maybe just enabling github pages and adding a link to the example in the README will avoid annoying people like me to report off topic issues.. ;)

Sorry for the large comment, ignorance and noise. Close this if you feel like (or I will in the short term) and thanks, but thanks a lot , for providing JS APIs on this topic!