Article
Open Access
Digital advertising fraud prediction using OLS regression
1 Faculty of Accountancy and Management, Universiti Tunku Abdul Rahman. Sungai Long Campus, Jalan Sungai Long, Bandar Sungai Long Cheras 43000, Kajang, Selangor, Malaysia
2 Faculty of Computing, Universiti Teknologi Malaysia, N28A, Lingkaran Ilmu, Johor Bahru 81310, Johor, Malaysia
Abstract
Digital advertising has become an essential tool for every business. The digital advertising budget has been increasing over the years. More businesses have turned to digital advertising during the pandemic. However, beginner advertisers might lack knowledge of digital advertising and, at the same time, pour extra capital into it. Furthermore, fraudulent activity in digital advertising is also increasing, which harms the current digital marketing environment, as well as every party involved. This study examines the factors that affect conversion fraud in digital advertising. A sample of 956 observations of computed-generated data is used to examine the variables related to conversion fraud. The results show that advertisers, ad logs, items, goals, and ad slots have a positive relationship with conversion fraud, which these variables determine the digital advertisement fraud. The predicted value for conversion fraud is 0.2936, which 280 of the samples are predicted to be fraud digital advertisement. Implications and recommendations for this research were discussed to facilitate the advertisers in future advertisement placement.
Keywords

digital advertising; ordinary least squares (OLS); clients; ad slot ID; item; goal; ad log type; pricing type; look up form

Preview