Recommendation System for E-commerce Based on Clustering and Association Rules
Download as PDF
DOI: 10.23977/iemss.2018.91452
Corresponding Author
Qu Shixin
ABSTRACT
[Objective]To generate recommendations based on customers’ behavior history and to evaluate it preliminary.[Methods] This paper clusters customers and uses the result to generate association rules. Then it calculates customers’ interests on each commodity. Finally the recommendation results are obtained. [Results]The performance of this method is pretty good when we use date from ‘Taobao’. It indicates that this idea is correct and it is significant for us to improve it later.[Limitations] This paper did not make good use of all data and the result of clustering and association rules should be improved.[Conclusions]This method can generate desirable recommendation results and it is meaningful to improve it in later research.
KEYWORDS
Personalized Recommendation Customers Clustering Association Rules