JavaScript-based machine learning system

I like to work based on web apps. It is very attractive to implement an application that can be run from anywhere and on any device. Over the past few months, I've been trying to get some basic lightweight machine learning algorithms that run on JavaScript and then use them to build " smart " Web apps. With the advent of node, it is possible to train the model on the server side and then use these models to make predictions on the client. I've researched some class libraries (tools) that do help with JavaScript-based machine learning. Now let's see what kind of libraries we have:

• Brain.js Neural Network Class Library
• SVMJS support Vector Machine class Library
• Forest.js implementation of random tree JavaScript class Library
• Numeric.js running Complex numerical calculations
• Node-sylvester Linear Algebra Class Library
• Lineareg.js linear regression class library based on JavaScript

Most of these libraries use node. JS to train the model on the server side. They (usually in JSON) generate a classifier that can make predictions on the server side. Generally, the generated classifiers are generally too cumbersome to be useful, but they are not useless. I hope I can build a simple and elegant application in the near future.

JavaScript-based machine learning system