TY - GEN
T1 - Intelligent mining on purchase information and recommendation system for e-commerce
AU - Xue, Weikang
AU - Xiao, Bopin
AU - Mu, Lin
N1 - Publisher Copyright:
© 2015 IEEE.
PY - 2016/1/18
Y1 - 2016/1/18
N2 - As an important marketing tool, recommendation systems for e-commerce offer an opportunity for merchants to discovery potential consumption tendency. This paper puts forward a novel recommendation algorithm to make the recommendation system more accurate, personalized and intelligent. Firstly, we use intelligent mining on purchase information, and regress consumer preference rating on click behavior. Secondly, we use Bipartite Network Recommendation model based on resource allocation and improved collaborative filtering model; the former abstracts products and consumers into nodes in the graph, and finds the correlation of products that recommend to others using alternative relation; and the latter solves the problem, caused by sparse data, by compressing rating matrix and predicting null values. Finally, according to Alibaba e-commerce customers purchase data, we verify that Hybrid Recommendation Model optimizes the accuracy and coverage of the recommendation results.
AB - As an important marketing tool, recommendation systems for e-commerce offer an opportunity for merchants to discovery potential consumption tendency. This paper puts forward a novel recommendation algorithm to make the recommendation system more accurate, personalized and intelligent. Firstly, we use intelligent mining on purchase information, and regress consumer preference rating on click behavior. Secondly, we use Bipartite Network Recommendation model based on resource allocation and improved collaborative filtering model; the former abstracts products and consumers into nodes in the graph, and finds the correlation of products that recommend to others using alternative relation; and the latter solves the problem, caused by sparse data, by compressing rating matrix and predicting null values. Finally, according to Alibaba e-commerce customers purchase data, we verify that Hybrid Recommendation Model optimizes the accuracy and coverage of the recommendation results.
KW - collaborative filtering
KW - purchase information
KW - recommendation system
KW - sparse data
UR - https://www.scopus.com/pages/publications/84962004057
U2 - 10.1109/IEEM.2015.7385720
DO - 10.1109/IEEM.2015.7385720
M3 - 会议稿件
AN - SCOPUS:84962004057
T3 - IEEE International Conference on Industrial Engineering and Engineering Management
SP - 611
EP - 615
BT - IEEM 2015 - 2015 IEEE International Conference on Industrial Engineering and Engineering Management
PB - IEEE Computer Society
T2 - IEEE International Conference on Industrial Engineering and Engineering Management, IEEM 2015
Y2 - 6 December 2015 through 9 December 2015
ER -