Prediction of Purchase Intention on the E-Commerce Clickstream Data

Creative Commons License

Gürbüz A., AKTAŞ M. S.

27th Signal Processing and Communications Applications Conference, 24 - 26 April 2019 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/siu.2019.8806311
  • Keywords: Clickstream data, Web usage mining, Incremental clustering, Anomaly Detection
  • Yıldız Technical University Affiliated: Yes


Every day millions of customers do shopping through e-commerce web sites around the world. Huge amounts of click stream data are generated from customers' use of e-commerce websites. Learning the behavior of users from click data and estimating the intention to buy has become an important need. Within the scope of this research, a methodology is proposed on the estimation of the purchase intention before finalizing the sessions of the customers. The proposed methodology was tested by the ACM RecSys Symposium on e-commerce data set published in 2015 and it was found that successful results could be obtained.