DMCA
Heterogeneous Graph-Based Intent Learning with Queries, Web Pages and Wikipedia Concepts
| Citations: | 1 - 0 self |
BibTeX
@MISC{Ren_heterogeneousgraph-based,
author = {Xiang Ren and Yujing Wang and Xiao Yu and Jun Yan and Zheng Chen and Jiawei Han},
title = {Heterogeneous Graph-Based Intent Learning with Queries, Web Pages and Wikipedia Concepts},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
The problem of learning user search intents has attracted intensive attention from both industry and academia. How-ever, state-of-the-art intent learning algorithms suffer from different drawbacks when only using a single type of da-ta source. For example, query text has difficulty in distin-guishing ambiguous queries; search log is bias to the order of search results and users ’ noisy click behaviors. In this work, we for the first time leverage three types of objects, namely queries, web pages and Wikipedia concepts collaboratively for learning generic search intents and construct a hetero-geneous graph to represent multiple types of relationships between them. A novel unsupervised method called hetero-geneous graph-based soft-clustering is developed to derive an intent indicator for each object based on the constructed het-erogeneous graph. With the proposed co-clustering method, one can enhance the quality of intent understanding by tak-ing advantage of different types of data, which complement each other, and make the implicit intents easier to interpret with explicit knowledge from Wikipedia concepts. Experi-ments on two real-world datasets demonstrate the power of the proposed method where it achieves a 9.25 % improve-ment in terms of NDCG on search ranking task and a 4.67% enhancement in terms of Rand index on object co-clustering task compared to the best state-of-the-art method.
Keyphrases
wikipedia concept web page heterogeneous graph-based intent learning real-world datasets search log first time leverage implicit intent user search intent generic search intent intensive attention intent understanding novel unsupervised method click behavior rand index explicit knowledge intent indicator het-erogeneous graph state-of-the-art method object co-clustering task hetero-geneous graph-based soft-clustering da-ta source distin-guishing ambiguous query single type multiple type different type different drawback hetero-geneous graph search result query text co-clustering method state-of-the-art intent tak-ing advantage
