Pemodelan dengan Generalized Lambda Distribution pada Data Kinerja Keuangan Perusahan Indonesia

Abstract

The generalized lambda distribution (GLD) is one of the statistical distributions that can be used for data matching. GLD is known to have a high level of flexibility so that it can create distributions with a variety of different forms. Even though GLD is known to have a high level of flexibility so that it can create distributions with a variety of different shapes, there are still gaps that need to be corrected in terms of obtaining initial values in order to ob- tain more optimal optimization results. In addition, in the numerical calcula- tion process, GLD uses a quasi-random number generator and the Nelder-Mead Simplex optimization algorithm, where there is still the possibility of using cer- tain random number generator methods, and alternative optimization algorithms other than Nelder-Mead Simplex to obtain more optimal results. . In this re- search, GLD was modified using a quasi-random number generator method with a Monte Carlo simulation approach, as well as simulating various parameter values to obtain more optimal model fit accuracy results. In the simulation pro- cess using data on the financial performance of companies listed on the Indone- sia Stock Exchange (IDX). The calculation simulation results can be accessed at https://gldapp.shinyapps.io/gldapp/. Meanwhile, financial performance data can be accessed at https://indcomp-stats.id/.