Abstract: As an important tool in correlation analysis, correlation coefficients have been extensively studied and applied in many science and engineering fields. Based on two commonly used bivariate contaminated Gaussian models, this paper reviews and compares the robustness of five correlation coefficients in environments with single-channel and double-channel impulsive noise, respectively. Theoretical results indicate that the most popular Pearson′s Product Moment Correlation Coefficient is very sensitive to impulsive noise interference. On the other hand, the other four coefficients demonstrate their robustness against impulsive noise in the two models.

Key words: Pearson′s product moment correlation coefficient(PPMCC), Spearman′s rho(SR), Kendall′s Tau(KT), Gini correlation (GC), Pearson′s rankvariate correlation coefficient(PRVCC), impulsive noise; , contaminated Gaussian model (CGM)