|Friday, February 21|
|CS07 Big Data in the Real World||
Fri, Feb 21, 11:00 AM - 12:30 PM
Using the Open Source R Language to Model Store Sales (302767)*John V. Colias, Decision Analyst, Inc.
Keywords: Predictive Analytics, R Language, Cross-Validation
A case study will demonstrate the use of R language predictive modeling to predict retail store sales, including examples of R code. A variety of data combined from multiple sources is used, including store performance metrics, trade area characteristics, competitive variables, and economic data. The case study will cover methodology for model training and validation, R packages used, comparison of relative accuracy of alternative types of predictive models (recursive partitioning, linear regression, random forest), and GIS mapping of results.