Aquaboutic | Focus Security Research | Vulnerability Exploit | POC


classic machine learning source code base (very complete, data mining, computer vision, pattern recognition, information retrieval related fields are applicable)

Posted by bax at 2020-04-14

Today I would like to introduce the classic open source machine learning software:

Programming language: the experimenter thinks that MATLAB is the most flexible (but the original version is very expensive), but the more promising is python (numpy + SciPy + Matlab) and C / C + +, so the combination can be used for research and commercial development, and the ease of use is not worse than matlab, and the function combination is more powerful. In my opinion, of course, R and Java are also good

1. Machine learning open source software network (including all kinds of machine learning programming languages, academic and commercial open source software)

2. Machine learning resource network occasionally found: (also very complete, 1 and 2 basically contain all ml classic open source software)

3 libsvm (the best in the field of support vector machine, needless to say, the masterpiece of Professor Lin from Taiwan University)

4 Weka (the most comprehensive and easy-to-use open source software based on Java machine learning algorithm)

Scikit (my favorite Python based machine learning software, the code is very well written, and the official documents are very complete, all have examples, algorithms are complete, and the development is also active, which is highly recommended for everyone to use)

Opencv (the most powerful open source computer vision library, with unlimited future, it must be used for image processing and pattern recognition. It can't be used to do experiments with MATLAB all day and disconnect with the industry, but it's difficult)

Orange (machine learning software based on C + + and python interface, beautiful interface, easy to call, can learn C + + and python at the same time, as well as visual functions,)

8 mallet (machine learning library based on Java, mainly used in natural language processing, characterized by good performance of Markov model and random domain, which can complement Weka)

9 nltk (Python's natural processing open source library, very easy to use and powerful, as well as several classic orelly tutorials)

Lucene (based on Java, including nutch, Solr, Hadoop, mahout, etc., it is an open source software that must be learned by comrades who are engaged in information retrieval and search engines, and must be learned by Java)

Of course, there are a lot of good open source software, which will be added one after another, to be continued, sleepy.....