Predictive Modeling for the Bid Responses in the Huvle

• Published in: Industrial Engineering & Management Systems Vol. 22; no. 2; pp. 144 - 152
• Main Authors: Cheon, Yu-Jeong, Hwang, Su-Hwan, Seo, Dong-Gil, Hwang, Go-Yeon, Hwang, Wook-Yeon
• Format: Journal Article
• Language: English
• Published: 대한산업공학회 2023
• Subjects: 산업공학; Demand Side Platform (DSP); Predictive Modeling; Supply Side Platform (SSP); Ad Exchange (AdX); Real-time Bidding (RTB); Machine Learning
• ISSN: 1598-7248; 2234-6473
• DOI: 10.7232/iems.2023.22.2.144

As an online advertising business emerges as an important tool for finding new customers and ending up with a profit creation, the real time bidding (RTB) system has gained a popularity in recent years. By connecting all the on-line advertisers and publishers, the RTB system facilitates the transactions of the advertising impressions based on the real-time auction. Here, an ad exchange (AdX) plays an indispensable role as an agent between the supply side plat-form (SSP) and the demand side platform (DSP) in the RTB system. Based on the RTB system, this research focuses on a perspective of the SSP which helps the publishers sell their impressions effectively. In the real-time auction, the SSP is required to send bid requests into AdXs which seek bidders who desire to buy the impressions. Therefore, it is crucial to predict which AdXs response the bid requests. Therefore, the bid response predictive models are developed in this research. In particular, we propose a novel performance evaluation approach for the bid response prediction based on the response rate and the F1 score. As a result, the bid response predictive models effectively reduce unnec-essary bid requests. KCI Citation Count: 0