Abstract

One of the important problems in modeling the data is the identification of suitable statistical model that fits the observed data. Many a time, the choice depends on the literature, type of distribution parameters that are needed, some specific distributional properties that the distribution should have etc. We consider four probability distributions to model skewed and leptokurtic data, and use the ratio of maximized likelihood statistic (RML) and the Kolmogorov-Smirnov (KS) distance to discriminate among the various overlapping family of distributions. Out of the probability models considered, we found that both RML and KS distance supports Log-logistic and Cauchy as a suitable fit for the stock prices. We also conduct a Wald test for samples from Log-logistic distribution.

Author: Mahesh K C, Ayush Jain, Kumar Tanna

Received on: February, 2024

Accepted on: December, 2024