 |
|
|||
Using the Bayesian Technique to the Tobit Quantile Regression with Scale Mixture Uniform for Selecting Important Variables Affecting Iraqi Investment Banking
Keywords
Tobit quantile regression, scale mixture uniform, asymmetric Laplace distribution, Investment banking
Abstract
High bank investment is important for supporting its funding center. But not all banks are active in investment banking, some banks its works limited on deposits and some banking service. In this paper, investment banking is a censored response variable at zero point. Therefore, the Tobit regression model (ToReg model) is considered more adaptable with censored response variable. But the ToReg model is no longer sufficient for data under study because (a) the distribution of investment banking shows skewness, (b) investment banking data have large gap between its small value and its large value, so this dataset has outlier values to overcome these problems by using the Tobit quantile regression model. Additionally, it gives us a complete description of the relationship between investment banking and a set of factors.