Improvement of Similarity Measure: Pearson Product-Moment Correlation Coefficient
Yongsuo Liu, Meng Qing-hua, Rong Chen, Jiansong Wang, Shumin Jiang, Yuzhu Hu · Journal of Chinese Pharmaceutical Sciences · 2004
Aim To study the reason of the insensitiveness of Pearson product-moment correlation coefficient as a similarity measure and the method to improve its sensitivity. Methods Experimental and simulated data sets were used. Results The distribution range of the data sets influences the sensitivity of Pearson product-moment correlation coefficient. Weighted Pearson product-moment correlation coefficient is more sensitive when the range of the data set is large. Conclusion Weighted Pearson product-moment correlation coefficient is necessary when the range of the data set is large.