Abstract
The paper reports our system in INEX 2014 Social Book Search (SBS) track. This is the second time that we attend the SBS track. Based on our social feature re-ranking system [1], we improve our system by involving some knowledge on understanding the queries. Our baseline system is built on Lucene [6], an open source information retrieval system. The new modification is a set of rules that can filter out unnecessary books from the recommendation list. The official run results show that the system performance is much improved than the 2013 system.
Original language | English |
---|---|
Pages (from-to) | 525-535 |
Number of pages | 11 |
Journal | CEUR Workshop Proceedings |
Volume | 1180 |
State | Published - 2014 |
Event | 2014 Cross Language Evaluation Forum Conference, CLEF 2014 - Sheffield, United Kingdom Duration: 15 Sep 2014 → 18 Sep 2014 |
Keywords
- Query type recognition
- Social Book Search
- Social features