Changing perspectives: Using graph metrics to predict purchase probabilities. (15th March 2018)

• Record Type: Journal Article
• Authors: Baumann, Annika; Haupt, Johannes; Gebert, Fabian; Lessmann, Stefan
• Is Part Of: Expert systems with applications. Volume 94(2018)
• Page Start: 137
• Page End: 148
• Publication Date: 2018-03-15
• Subjects: Predictive analytics -- Clickstream data -- User graph -- Graph metrics; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2017.10.046
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: We assess the applicability of graph metrics to predict purchase probabilities. Real-world clickstream data of two online retailers is used. Graphs are derived out of sessions of website visitors. Distance- and centrality-based graph metrics are useful for prediction. Closeness vitality, radius, number of circles and self-loops are most important. Abstract: The prediction of online user behavior (next clicks, repeat visits, purchases, etc.) is a well-studied subject in research. Prediction models typically rely on clickstream data that is captured during the visit of a website and embodies user agent-, path-, time- and basket-related information. The aim of this paper is to propose an alternative approach to extract auxiliary information from the website navigation graph of individual users and to test the predictive power of this information. Using two real-world large datasets of online retailers, we develop an approach to construct within-session graphs from clickstream data and demonstrate the relevance of corresponding graph metrics to predict purchases.