The MM Algorithm Based On The Quantile Method

With the development of statistics, quantile in the economic, environmental science, survival analysis has received a wide range of applications, the calculation of quantile is particularly important.Known general distribution function is present and only when the overall capacity is small, can take sort of way obtained some points of the overall site, when the overall capacity of large, ordering method becomes complicated.MM algorithm is an alternative to the L1 objective function by constructing a new objective function Q(?)(μ│μk)optimization problem to find solutions of iterative algorithm.This article discusses the use of MM algorithms for non-classified sample quantiles problem; And proposed the use of random sampling methods to address the huge amount of data quantile calculation, calculation of excessive and time-consuming issues such as the algorithm.