EDSIGCON Proceeding 2015

Wilmington, North Carolina

2015 EDSIG Proceedings - Abstract Presentation

Teaching Business Analytics Using the caret Package in R

Kelley Donalds
Bridgewater State University

R is a powerful, open source statistical programming environment that is becoming increasingly popular within the business community. The College of Business at Bridgewater State University recently piloted an “Introduction to Business Analytics using R” course which had only an introductory statistics course prerequisite. A major challenge of teaching such a course is covering R programming, new analytic techniques and the data analytics process. To meet this challenge, one approach is to focus on R programming by teaching students how to do tasks that they have already learned. This would include, for example, conducting t-tests and analyses of variance. A second approach is to focus on new analytical techniques and to minimize the burden of learning R by providing sample scripts that students can easily adapt for assignment completion. However, a drawback of the second approach which was used in the pilot course is that it is easy for students to become overly preoccupied with the details of particular implementations of particular techniques. A third approach currently under development is to focus on the data analytics process by relying on the caret package in R (Kuhn 2008). Caret provides a common interface to many predictive techniques and automates the process of data preparation, model building and model validation. This third approach has the advantage of providing a “big picture” perspective that focuses on tasks common to all predictive modeling projects and also illustrates the power of open source programming environments. In particular, the caret package is itself an example of how existing software developed by many contributors can be reused and combined to provide novel capabilities. In this presentation, I will discuss this third approach in more detail and describe its benefits from a pedagogical perspective.

References: Kuhn, Max (2008). “Building Predictive Models in R Using the caret Package.” Journal of Statistical Software, 28(5), URL: http://www.jstatsoft.org/v28/i05/paper

Recommended Citation: Donalds, K., (2015). Teaching Business Analytics Using the caret Package in R. Proceedings of the EDSIG Conference, (2015) n.3607, Wilmington, NC