Beijing, China
Beijing, China

Time filter

Source Type

Zhu Y.-X.,University of Electronic Science and Technology of China | Lu L.-Y.,Hangzhou Normal University | Lu L.-Y.,University of Fribourg | Lu L.-Y.,Baifendian
Dianzi Keji Daxue Xuebao/Journal of the University of Electronic Science and Technology of China | Year: 2012

In this article, the existed evaluation metrics for recommender systems are reviewed and the new progresses in this field are summarized from four aspects: accuracy, diversity, novelty and coverage. The merits, weaknesses and applicable conditions of different evaluation metrics are analized. The focus is concentrated on the importance of rank and some representative rank-sensitive metrics. The user-centric recommender systems are discussed and some important open problems are outlined as future possible directions.


Liu J.-H.,University of Electronic Science and Technology of China | Zhou T.,University of Electronic Science and Technology of China | Zhang Z.-K.,Hangzhou Normal University | Yang Z.,University of Electronic Science and Technology of China | And 4 more authors.
PLoS ONE | Year: 2014

As one of the major challenges, cold-start problem plagues nearly all recommender systems. In particular, new items will be overlooked, impeding the development of new products online. Given limited resources, how to utilize the knowledge of recommender systems and design efficient marketing strategy for new items is extremely important. In this paper, we convert this ticklish issue into a clear mathematical problem based on a bipartite network representation. Under the most widely used algorithm in real e-commerce recommender systems, the so-called item-based collaborative filtering, we show that to simply push new items to active users is not a good strategy. Interestingly, experiments on real recommender systems indicate that to connect new items with some less active users will statistically yield better performance, namely, these new items will have more chance to appear in other users' recommendation lists. Further analysis suggests that the disassortative nature of recommender systems contributes to such observation. In a word, getting in-depth understanding on recommender systems could pave the way for the owners to popularize their cold-start products with low costs. Copyright: © 2014 Liu et al.


Zhang L.,Baifendian | Zhang L.,University of Electronic Science and Technology of China | Bai L.-S.,Baifendian | Zhou T.,University of Electronic Science and Technology of China | Zhou T.,Hangzhou Normal University
Dianzi Keji Daxue Xuebao/Journal of the University of Electronic Science and Technology of China | Year: 2013

Personalized recommendation has now been widely used in E-commerce, but there are still some problems to be solved such as cold-start problem, data sparsity, diversity-accuracy dilemma and so on. Existing literatures have focused on single data set, lacking a systematic understanding about the accessing behavior involving multiple web sites. Thanks to the real data, provided by Baifendian Information Technology recommendation engine, we analyze users' behavior on multi-B2Cs (business-to-customers) and propose a crossing recommendation algorithm which is able to recommend items of a B2C site to users according to the records of users in other B2C web sites. This algorithm largely improves accuracy compared with purely random recommendation under completely cold-start environment and can still keep high diversity and novelty.


Baifendian | Entity website

1 2 3 4 1


Baifendian | Entity website

2016-05-24


Baifendian | Entity website

1542% A4313LED1LCD445995199549959997999ALED A6A6 APP


Baifendian | Entity website


Baifendian | Entity website

2000


Baifendian | Entity website


Loading Baifendian collaborators
Loading Baifendian collaborators