An Empirical Study on Performance Measures for Online Advertising

Author

McInroy, Ben ; Wenying Feng ; Yuanyi Pan

Author_Institution

Depts. of Math. & Comput. & Inf. Syst., Trent Univ., Peterborough, ON, Canada

fYear

2015

fDate

13-15 April 2015

Firstpage

38

Lastpage

43

Abstract

In this paper, we study performance measures for online advertising systems by empirical approach. The traditional statistical measure Critical Success Index (CSI), also known as Threat Score (TS), is modified as CSI-R to add the emphasis on recall. We then generate two realistic data sets to examine the effects of CSI and CSI-R in evaluation of a Real-Time Bidding (RTB) system for online advertising. Performance factors such as accuracy, lift, efficiency and skill score are compared for both minority and majority classes. The results show that CSI-R is better applied on measuring such a system.

Keywords

Internet; advertising; CSI-R; RTB system; TS; online advertising systems; performance measures; real-time bidding; skill score; statistical measure critical success index; threat score; Accuracy; Advertising; Data models; Digital signal processing; Indexes; Optimization; Predictive models; Accuracy; Critical Success Index (CSI); Real-Time Bidding (RTB); lift; online advertising; performance measure; recall; skill score;

fLanguage

English

Publisher

ieee

Conference_Titel

Information Technology - New Generations (ITNG), 2015 12th International Conference on

Conference_Location

Las Vegas, NV

Print_ISBN

978-1-4799-8827-3

Type

conf

DOI

10.1109/ITNG.2015.12

Filename

7113446