Transfer Learning for Content-Based Recommender Systems Using Tree Matching

Abstract : In this paper we present a new approach to content-based transfer learning for solving the data sparsity problem in cases when the users’ preferences in the target domain are either scarce or unavailable, but the necessary information for the preferences exists in another domain. Training a system to use such information across domains is shown to produce better performance. Specifically, we represent users’ behavior patterns based on topological graph structures. Each behavior pattern represents the behavior of a set of users, when the users’ behavior is defined as the items they rated and the items’ rating values. In the next step, a correlation is found between behavior patterns in the source domain and target domain. This mapping is considered a bridge between the two. Based on the correlation and content-attributes of the items, a machine learning model is trained to predict users’ ratings in the target domain. When our approach is compared to the popularity approach and KNN-cross-domain on a real world dataset, the results show that our approach outperforms both methods on an average of 83%.
Complete list of metadatas

Cited literature [12 references]  Display  Hide  Download

https://hal.inria.fr/hal-01506793
Contributor : Hal Ifip <>
Submitted on : Wednesday, April 12, 2017 - 11:19:19 AM
Last modification on : Wednesday, December 12, 2018 - 11:24:12 AM

File

978-3-642-40511-2_28_Chapter.p...
Files produced by the author(s)

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

  • HAL Id : hal-01506793, version 1

Citation

Naseem Biadsy, Lior Rokach, Armin Shmilovici. Transfer Learning for Content-Based Recommender Systems Using Tree Matching. 1st Cross-Domain Conference and Workshop on Availability, Reliability, and Security in Information Systems (CD-ARES), Sep 2013, Regensburg, Germany. pp.387-399. ⟨hal-01506793⟩

Share

Metrics

Record views

269

Files downloads

247