Shulou(Shulou.com)06/02 Report--

This article mainly explains "What are the machine learning libraries for JavaScript", interested friends may wish to take a look. The method introduced in this paper is simple, fast and practical. Let's let Xiaobian take you to learn "What are the machine learning libraries for JavaScript"!

Python is a general-purpose programming language used not only for machine learning but also for scientific computing, back-end Web development, desktop applications, etc. R is primarily used for statisticians. However, they share at least two characteristics:

They are suitable for non-programmers

They have a comprehensive ML library

In many cases, ML algorithms are implemented in Fortran, C, C++, or Cython and invoked from Python or R.

Java is also used for machine learning, but is usually used by professional programmers.

JavaScript has gained popularity over the past few years and there are some very interesting machine learning libraries that can implement ML methods on browsers or Node.js. Surprisingly, many of these libraries implement a lot of code in JavaScript.

ml.js

ml.js is a comprehensive, generic JavaScript ML library for browsers and Node.js. It provides the following routines:

Bit operations on arrays, hash tables, sorting, random number generation, etc.

Linear algebra, array operations, optimization (Levenberg-Marquardt method), statistics

cross-validation

supervised learning

unsupervised learning

Supported supervised learning methods are:

Linear, polynomial, exponential and power regression

k-nearest neighbor

Naive Bayes

support vector machine

Decision Trees and Random Forests

Feedforward neural networks, etc.

In addition, ml.js provides several unsupervised learning methods:

principal component analysis

Cluster analysis (k-means and hierarchical clustering)

Self-organizing map (Kohonen network)

TensorFlow.js

TensorFlow is one of the most popular machine learning libraries. It focuses on various types and structures of artificial neural networks, including deep networks as well as components of networks.

TensorFlow was created by Google Brain Team and written in C++ and Python. However, it can be used with many languages including JavaScript.

TensorFlow is a very comprehensive library that still makes it easy to build and train models. It supports a wide variety of network layers, activation features, optimizers and other components. It has good performance and offers GPU support.

TensorFlow.js is a JavaScript ML library for browsers or Node.js. It supports WebGL.

brain.js

brain.js is a library written in JavaScript-focused on training and applying feedforward and recurrent neural networks. It also provides other utilities, such as mathematical routines needed for neural networks.

It offers advanced options such as:

Training networks using GPUs

Asynchronous training that can accommodate multiple networks in parallel

Cross-validation is a more sophisticated validation method

brain.js saves or loads the model to or from a JSON file.

ConvNetJS

ConvNetJS is another library for neural networks and deep learning. It can train neural networks in the browser. In addition to classification and regression problems, it also has reinforcement learning modules (using Q-learning) that are still experimental. ConvNetJS supports convolutional neural networks that excel in image recognition.

In ConvNetJS, a neural network is a list of layers. It provides the following layers:

Enter (***) layers

fully connected layers

convolutional layer

collection layer

local contrast normalization layer

Classifier missing (output) layers: softmax and svm

Regression loss (output) layer using L2

It supports several important activation functions such as:

RELU

sigmoid colon

hyperbolic tangent

MAXOUT

and optimizers such as:

stochastic gradient descent

Adadelta

AdagradS

ConvNetJS also provides a convenient way to save and load models of JSON files.

License: Massachusetts Institute of Technology.

WebDNN

WebDNN is a library focused on deep neural networks, including recurrent neural networks with LSTM architecture. It is written in TypeScript and Python and provides JavaScript and Python APIs.

It also offers the possibility of executing GPU in the browser.

A very handy feature of WebDNN is the ability to convert and use PyTorch, TensorFlow, Keras, Caffemodel or Chainer pre-trained models.

natural

natural is a JavaScript library for natural language processing of Node.js.

It supports:

tokenization (breaking text into an array of strings)

Calculation of Chord Distance

Match similar strings

Classification (naive Bayes, logistic regression and *** entropy)

Sentiment analysis (currently available in eight languages)

Speech matching, reflectors, n-gram, etc.

At this point, I believe that everyone has a deeper understanding of "what machine learning libraries are used for JavaScript", so let's actually operate it! Here is the website, more related content can enter the relevant channels for inquiry, pay attention to us, continue to learn!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.