5 ways to bring machine learning to programming languages like Java, Python, and go

This is a creation in Article, where the information may have evolved or changed.

5 ways to bring machine learning to programming languages like Java, Python, and go

Machine learning is hot, and this article collects common and useful open-source machine learning tools in programming languages such as Java, Python, and go, as well as developers interested in machine learning or data scientists who are prepared to deal with machine learning.

"Editor's note" machine learning seems to be the focus of the night from the unknown to the attention of the pawn, open-source tools for machine learning is also growing, but the current challenge is how to make machine learning interested in developers and ready to use its data scientists really use them, This article collects the common and useful open-source machine learning tools in several languages, which is very noteworthy, this article is from InfoWorld.

The following is the original text:

After decades of development as a professional discipline, machine learning seems to be in front of us all night as a business tool of the limelight. The challenge now is how to make it effective, especially for developers and the data scientists who are preparing to use it.

To this end, we have collected some of the most common and useful open-source machine learning tools, through this article to share with you.

Python

Data scientists are embracing python in the hope that there is another more open alternative to the R language, and many employers today are looking for big data experts, and Python is a necessary skill. As a result, a large number of machine learning repositories are beginning to appear in the ever-expanding list of Python software.

The first recommendation is Scikit-learn, which can be loaded into algorithms and modules, and is widely appreciated on GitHub (with the number of fork versions approaching 2000) and favored by many tech giants. Another closely followed is Pybrain, which is designed to reduce the difficulty of use and provide the ability to connect with other powerful tools. As the name implies, Pybrain's focus is on neural networks and unsupervised learning, and it also provides a set of mechanisms for training and redefining algorithms.

Go

Google's system language, due to its parallel design, makes it seem to be an ideal environment for writing machine learning libraries. Although the current library project is still small in size, there are some notable concerns, Golearn, which its developers describe as a "built-in Battery" machine learning library. It provides a variety of tools such as filtering, sorting, and regression analysis. Another set of smaller and more basic libraries is mlgo, although it currently offers a very small number of algorithms, but plans to launch more in the future.

Java on Hadoop

In the big data frame that people love, Mahout (meaning "Elephant Rider" in Hindi) contains several common machine learning methods. This package is built around algorithms rather than methods, so you need to have a certain algorithm base, in other words, if you're serious enough, it's certainly not difficult to see how the various parts of the functionality are integrated, for example, you can build a user-based referral system with a few lines of code.

Another Hadoop-based machine learning project is the Oryx of Cloudera, which is characterized by further analysis of mahout processing results by delivering live stream results rather than processing batch jobs. The project is still in its infancy, and note that this is a project rather than a real product, but it is constantly improving, so it deserves attention.

Java

In addition to the above-mentioned mahout for Hadoop, other Java-oriented machine learning libraries are also widely used. Weka is a desktop application developed by the University of Waikato in New Zealand, which adds visualization and data mining capabilities to common algorithm collections. For those who want to create a front end for their work or plan to use Java as the initial development, Weka may be the best choice. Java-ml is good, but it's better for developers who are used to working with Java and machine learning.

Javascript

The joke about JavaScript ("Blackwood's Law"), is that any content that can be written by JavaScript will eventually be written by JavaScript, which is also true for machine learning libraries. The number of scenarios currently written by JavaScript is still relatively small in this area, and most options are just a single algorithm rather than a complete library, but some useful tools have come to the fore. Convnetjs allows for deep-learning neural network training directly in the browser, while the name Brain provides the neural network as an installable NPM module for everyone. In addition, the ENCOG library is also noteworthy, and it works on a variety of platforms: Java, C #, C + +, and JavaScript.