A Novel Data Statistical Analysis Method Based on Hyper-Erlang Distribution

Selecting the appropriate probability distribution can be a challenging task when using parametric methods for statistical analysis of specific data. To address this challenge, this paper proposes a method of analyzing the data using the Hyper-Erlang distribution. This distribution is a dense subset of all probability distributions on the non-negative real axis and is both easy to analyze and compute. This paper discusses the dense properties of the Hyper-Erlang distribution and the parameter estimation method based on the expectation maximization algorithm, and verifies the validity of the proposed method using the actual income data.