Maximal information coefficient (MIC) maximum mutual information coefficient

Mic is a method used to calculate the correlation between variables in a large data set, which is effective: no assumptions are made on the distribution of data, and the correlations between variables (both linear and nonlinear) are estimated

Download the compressed file in the link below and unzip it to compile:

http://pan.baidu.com/s/1bSEG2a

1. Open matlab switch to the following directory

cd ..... \minepy-1.2.0\minepy-1.2.0\matlab

2. Compiling the mine.c file with the MEX directive

Mex MINE_MEX.C. /libmine/mine.c

3. The Mine_mex.mex (or Mine_mex.mexw32 file, depending on the system) will be obtained under the current directory (/matlab/)

4. Put the MINE.M and mine_mex.mexw32 files in the same folder to use

5. Official Examples:

Operation Result:

A mic value of 1.0 proves that there is a strong correlation between x and Y (the sinusoidal relationship is used in the example) and that the mic can detect nonlinear correlations

Note: X and y must be line vectors

PS: The installation package in the link can also be installed in Python, if the link fails, you can leave a message in the blog, and give the receiving mailbox.

Reference Links:

http://blog.csdn.net/u014271612/article/details/51781250