Session

Management, Business and Economics

Description

A model is a statement of reality or its approximation. Most phenomena in the social sciences are extremely complex. With a model we simplify the reality and focus on a manageable number of factors. It is impossible to completely understand why consumers default and identify all the factors influencing customer’s default behavior. The bank manager sets up a statistical model that relates customer’s default behavior to only two important factors, the income and the education. There surely are thousands of other variables that may influence customer’s default behavior. This article describes the fundamentals of a statistical model building. We begin our discussion on the managerial justification for building a statistical model. Then we discuss three important statistical issues that are of prime importance to database marketers: model/variable selection, treatment of missing data, and evaluation of the model using Bootstrap Method.

Keywords:

Predictive modelling, Customer, Evaluation, Bootstrap Method

Proceedings Editor

Edmond Hajrizi

ISBN

978-9951-550-12-3

First Page

93

Last Page

98

Location

Durres, Albania

Start Date

7-11-2015 9:00 AM

End Date

7-11-2015 5:00 PM

DOI

10.33107/ubt-ic.2015.28

Included in

Business Commons

Share

COinS
 
Nov 7th, 9:00 AM Nov 7th, 5:00 PM

Statistical Issues in Predictive Modelling using Bootstrap Method

Durres, Albania

A model is a statement of reality or its approximation. Most phenomena in the social sciences are extremely complex. With a model we simplify the reality and focus on a manageable number of factors. It is impossible to completely understand why consumers default and identify all the factors influencing customer’s default behavior. The bank manager sets up a statistical model that relates customer’s default behavior to only two important factors, the income and the education. There surely are thousands of other variables that may influence customer’s default behavior. This article describes the fundamentals of a statistical model building. We begin our discussion on the managerial justification for building a statistical model. Then we discuss three important statistical issues that are of prime importance to database marketers: model/variable selection, treatment of missing data, and evaluation of the model using Bootstrap Method.