Name
Abstract | ||
|---|---|---|
With the recent proliferation of e-commerce services, online shopping has become more and more popular among customers. Because it is necessary to recommend proper items to customers, to improve the accuracy of recommendation, high-performance recommender systems are required. However, current recommender systems are mainly based on information of their own domain, resulting in low accurate recommendation for customers with limited purchasing histories. The accuracy may suffer due to a lack of information. In order to use information from other domains, it is necessary to associate behaviors in different domains of the behaviorally related users. This paper presents a preliminary analysis of matching behaviors of the behaviorally related users in different domains. The result shows that we got a better prediction rate than linear regression.
|
Year | DOI | Venue |
|---|---|---|
2019 | 10.1145/3308560.3316595 | Companion Proceedings of The 2019 World Wide Web Conference |
Keywords | Field | DocType |
behavior matching, canonical correlation analysis, multi-domains | Recommender system,Data mining,Computer science,Canonical correlation,Purchasing,Artificial intelligence,Machine learning,Linear regression | Conference |
ISBN | Citations | PageRank |
978-1-4503-6675-5 | 0 | 0.34 |
References | Authors | |
0 | 7 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Yuan Lyu | 1 | 0 | 0.34 |
| Daichi Amagata | 2 | 43 | 13.26 |
| Takuya Maekawa | 3 | 326 | 49.93 |
| Takahiro Hara | 4 | 1819 | 193.85 |
| Hao Niu | 5 | 153 | 11.33 |
| Kei Yonekawa | 6 | 2 | 3.07 |
| Mori Kurokawa | 7 | 34 | 5.59 |