Abstract
In order to measure whether and how things are related, statistical correlation analysis comes into being. Among them, Pearson coefficients, Spearman and Kendall coefficients are widely used, but these correlation analysis methods cannot detect a wide range of relationship types due to their own limitations. Therefore, in 2011, Reshef et al. introduced a new correlation analysis method, maximal information coefficient (MIC), but this method has low computational efficiency and does not give an optimal data division method. On this basis, we propose the maximum information coefficient based on K-Medoids clustering (KM-MIC), which combines the K-Medoids clustering algorithm to optimize the way of data partitioning, and can quickly calculate whether there is correlation between the data. And this method has two major characteristics, generality and equitability.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
