Visualization and Statistical Modeling of Financial Big Data: Log-Linear Modeling With Skew Error

Abstract

In this paper, we consider the visualization and statistical modeling of financial data (e.g., sales, assets) for many global firms which are listed and delisted. This study presents an exploratory data analysis carried out in the R programming language. The results show that a log-linear model with skew-t error is useful for modeling the total sales volume (in thousands of U.S. dollars) as a function of the number of employees and the total assets (in thousands of U.S. dollars), and is obtained by comparing the Akaike information criteria between several log-linear models with error terms which are independent and identically distributed random variables with skew distributions. These models are also evaluated by cross-validation.