Online Program

Saturday, February 20
PS3 Poster Session 3 & Continental Breakfast sponsored by Capital One Sat, Feb 20, 8:00 AM - 9:15 AM
Ballroom Foyer

Combining Multiple Statistical Methods for Measuring Ad Effectiveness (303237)

*Buck Fisk, Interspace co ltd 
Mizuki Hinata, Interspace co ltd 

Keywords: R, inference, survival analysis, small data, bootstrap, online ads

Data mining and business analytics involve unconventional and complex data that do not lend to one or two focused statistical methods. Taking a more broad approach as opposed to a few focused methods is better suited for the exploratory nature of mining unconventional data. Here, we analyze a single set of online advertising data by implementing multiple statistical methods such as survival analysis, bootstrapping, principal component analysis, logistic regression, and clustering. Then, we apply the same methods while withholding specific data to examine the impact on the results. By doing so, we better understood which statistical methods provide practical support for our clients and internal operations.