The Numerical Analysis of KL Quantile Estimates

ABSTRACT

This article uses R software KL quantile estimate numerical simulation under different distributions and different sample values, and on this basis, two cases of KL SQ sample quantile quantile estimates and estimated mean square numerical errors, the simulation results show that: in most cases, KL quantile estimate of the mean square error is less than SQ sample quantile estimation mean square error; the truncated distribution (such as exponential distribution and uniform distribution) cut end of the estimated effect of KL quantile estimate is very good, and much better than the median estimate SQ sample points; heavy tail of the distribution (e.g. distribution) have an impact on the estimated effects of KL quantile estimates.

Keywords: