Nonparametric Multiple Comparison Test for Change-Point Detection in Big Data

Abstract

We offer a new effective tool for detection of change-points in tracking data (movement data, health rate data etc.). We developed a nonparametric test for homogeneity of data in two adjacent time intervals. In the context of Big Data and IoT it allows online analyzing data stream from sensor and recognizing significant deviations from the baseline. The significance level for the test is less than 0.05. Also, we provide the results of comparison of the test with well-known Kolmogorov–Smirnov test, the sign test, the Wilcoxon signed-rank test, and the Mann–Whitney test. The computational experiment has shown that the Klyushin–Petunin test based on p-statistics has very high robustness, specificity, and sensitivity, and is more universal than the Kolmogorov–Smirnov test, the sign test, the Wilcoxon signed-rank test, and the Mann–Whitney test.