Note on the comparison of some outlier labeling techniques

Abstract

Problem statement: Methods proposed for estimating and resolving outliers are compared. Approach: In this respect, we exploit three well-known classifiers for identifying outliers to establish guidelines for the choice of outlier detection methods. Results: It was shown that the standard deviation is inappropriate to use here because it is highly sensitive to extreme values. Conclusion/Recommendations: The result of these estimated outliers is a better way of resolving large population. © 2011 Science Publications.