The School of Computer Science is pleased to present…
Using Sequential Multi-Behaviour Product Features for E-commerce Recommendation
MSc Thesis Proposal by: Saadhika Bandreddy
Date: Friday December 15th, 2023
Time: 2:00 pm to 4:00 pm
Location: Essex Hall, Room 122
In most real-world recommender systems, users interact with items in a sequential and multi-behavioural manner. There are various types of user multi-behaviours in practical scenarios, for example, there are clicks, likes, add-to-cart, and purchases. Analyzing the fine-grained relationship of items behind the users’ multi-behaviour interactions is critical in improving the performance of recommender systems. Existing methods, such as DACBRec21, MBHT22, and HSPRec19 use the customer multi-behaviour information to improve the accuracy of recommendations. DCABRec21 system uses multiple user behaviours and negative feedback in Collaborative Filtering (CF) methods. HPCRec18 system used purchase frequency and consequential bond between clicks and purchased data to improve the user-item frequency matrix. MBHT22 systems is a multi-behaviour recommendation system that uses a hypergraph-transformer. The HSPRec19 system converts historic click and purchase data to sequential data and enhances the user-item frequency matrix with sequential purchase patterns. HSPRec19 system generates recommendations based on frequent sequential purchase patterns and does not capture the item-level multi-behaviour dependencies to alleviate to a larger extent, data sparsity problems.
This thesis proposes a system called the Multi-Behaviour Sequential Pattern Recommendation System (MBSPRec System), which is an extension of the HSPRec19 system that includes multi-behavior frequent patterns along with frequent click and purchase patterns to improve the quality of recommendations and alleviate data sparsity to a larger extent. The proposed MBSPRec generates a Multi-Behaviour Sequential Database for each user behaviour type using the Multi-Behaviour Sequential Database Generator (MBSDBG) and Multi- Behaviour Sequential Pattern Miner (MBSPM) that mines multiple user behaviour sequential pattern rules that yield additional sequential patterns to overcome data sparsity of User-Item Matrix and improve the accuracy of the recommendations. The proposed MBSPRec mines approximate sequential data using the ApproxMAP algorithm to improve the Consequential Bond between multiple behaviour and purchase sequences to give multi-behaviour frequent sequential rules where no purchase has happened.
Keywords — Data mining, Sequential pattern mining, Collaborative filtering, approximate patterns, Multi-behaviour Recommender system, E-commerce Recommender systems, Data sparsity
Internal Reader: Dr. Ahmad Biniaz
External Reader: Dr. Dennis Borisov
Advisor: Dr. Christie Ezeife