Correlation quantifies the relationship between features in order to identify feature candidates that may be best suited to achieve desired effects. Linear correlation methods are robust and computationally efficient but detect only linear dependencies. Nonlinear correlationmethods are able to detect nonlinear dependencies but need to be carefully parametrized. As a popular example for nonlinear correlation we present the chi-square test for independence that is based on histogram counts. Nonlinear correlation can also be quantified by the regression validation error. Correlation does not imply causality, so correlation analysismay reveal spurious correlations. If the underlying features are known, then spurios correlations may be handled with partial correlation methods.
KeywordsCorrelation Method Causal Connection Nonlinear Dependency Spurious Correlation Feature Candidate
Unable to display preview. Download preview PDF.
- 1.D. Freedman, R. Pisani, and R. Purves. Statistics. W. W. Norton & Company, New York, 2007.Google Scholar
- 3.T. A. Runkler. Fuzzy histograms and fuzzy chi–squared tests for independence. In IEEE International Conference on Fuzzy Systems, volume 3, pages 1361–1366, Budapest, July 2004.Google Scholar