A novel statistical maximal information coefficient (MIC) that can detect the nonlinear relationships in large data sets was proposed by Reshef et al. (2011), with emphasis being placed on the equitability, which is a very important concept in data exploration. In this paper, an improved algorithm for approximation of the MIC (IAMIC) is proposed for the development of the equitability. Based on quadratic optimization processes, the IAMIC can search for a more optimal partition on the y-axis rather than use that which was obtained simply through the equipartition of the y-axis, to enable it to come closer to the true value of the MIC. It has been proved that the IAMIC can search for a local optimal value while using a lower number of iterations. It has also been shown that the IAMIC provides higher accuracy and a more acceptable run-time, based on both a mathematical proof and the results of simulations.
|Number of pages||18|
|Journal||International Journal of Data Warehousing and Mining|
|Publication status||Published - 1 Jan 2017|
- Big data
- Quadratic optimization
ASJC Scopus subject areas
- Hardware and Architecture