ecms_neu_mini.png

Digital Library

of the European Council for Modelling and Simulation

 

Title:

Classification Of E-Customer Sessions Based On Support Vector Machine

Authors:

Grazyna Suchacka, Magdalena Skolimowska-Kulig, Aneta Potempa

Published in:

 

 

(2015).ECMS 2015 Proceedings edited by: Valeri M. Mladenov, Grisha Spasov, Petia Georgieva, Galidiya Petrova, European Council for Modeling and Simulation. doi:10.7148/2015

 

 

ISBN: 978-0-9932440-0-1

 

29th European Conference on Modelling and Simulation,

Albena (Varna), Bulgaria, May 26th – 29th, 2015

 

Citation format:

Grazyna Suchacka, Magdalena Skolimowska-Kulig, Aneta Potempa (2015). Classification Of E-Customer Sessions Based On Support Vector Machine, ECMS 2015 Proceedings edited by: Valeri M. Mladenov, Petia Georgieva, Grisha Spasov, Galidiya Petrova  European Council for Modeling and Simulation. doi:10.7148/2015-0594

 

DOI:

http://dx.doi.org/10.7148/2015-0594

Abstract:

A key feature of high-traffic e-commerce sites is the ability to offer a predictive and personalized service to Web users. Visitors to online stores are potential buyers but in reality very few visits finally result in a product purchase. Thus, it would be especially valuable for online retailers to predict buyers against browsers based on some session features (e.g. session duration, the number of downloaded pages, the kind of realized Web interactions) and some HTTP-level information (the number of HTTP requests, the volume of data transfer in session). In this paper, we recast online purchase pre- dictions as a classification problem. Every user session in a web store is represented as a 23-element vector in the session feature space. Based on historical data from an online bookstore an SVM classification model is proposed, dividing user sessions into two classes: browsing sessions and buying sessions. The best SVM classifier proved to be very effective, with a predictive accuracy of over 99% and the probability of predicting a buying session of almost 95%.

 

Full text: