Entity ranking using Wikipedia as a pivot

Abstract

In this paper we investigate the task of Entity Ranking on the Web. Searchers looking for entities are arguably better served by presenting a ranked list of entities directly, rather than a list of web pages with relevant but also potentially redundant information about these entities. Since entities are represented by their web homepages, a naive approach to entity ranking is to use standard text retrieval. Our experimental results clearly demonstrate that text retrieval is effective at finding relevant pages, but performs poorly at finding entities. Our proposal is to use Wikipedia as a pivot for finding entities on the Web, allowing us to reduce the hard web entity ranking problem to easier problem of Wikipedia entity ranking. Wikipedia allows us to properly identify entities and some of their characteristics, and Wikipedia's elaborate category structure allows us to get a handle on the entity's type.Our main findings are the following. Our first finding is that, in principle, the problem of web entity ranking can be reduced to Wikipedia entity ranking. We found that the majority of entity ranking topics in our test collections can be answered using Wikipedia, and that with high precision relevant web entities corresponding to the Wikipedia entities can be found using Wikipedia's 'external links'. Our second finding is that we can exploit the structure of Wikipedia to improve entity ranking effectiveness. Entity types are valuable retrieval cues in Wikipedia. Automatically assigned entity types are effective, and almost as good as manually assigned types. Our third finding is that web entity retrieval can be significantly improved by using Wikipedia as a pivot. Both Wikipedia's external links and the enriched Wikipedia entities with additional links to homepages are significantly better at finding primary web homepages than anchor text retrieval, which in turn significantly improved over standard text retrieval.

Ranking Entities for Web Queries Through Text and Knowledge

“…However, alternative or broader definitions are possible (see, e.g., [2,13]). The focus on type queries led to the development of systems that leveraged Wikipedia category system -e.g., category overlap [32], or measures capturing the strength of association between terms and category labels [27]. Broadly similar in spirit to our WikiSDM approach, additional structured content from Wikipedia such as links can be also used to perform document retrieval [32,10].…”

A Novel Entity Type Filtering Model for Related Entity Finding

“…Three levels of relevance are examined which are document, passage and entity, respectively. R. Kaptein et al [11] propose an approach using Wikipedia as a pivot for finding entities on the web, reducing the hard web entity ranking problem to easier problem of Wikipedia entity ranking.…”

An Approach of Filtering Wrong-Type Entities for Entity Ranking

“…[3] proposed an approach which manually assigned a number of initial Wikipedia categories for given target-entity type and searched continuously the sub-categories of initial categories as the target Wikipedia categories. M. Koolen [2]'s method and R. Kaptein [4]'s method both have some shortcomings: (1) it need manual participation and a set of topics have same target Wikipedia categories, (2) the filtering model only has two values. R. Kaptein et al [4] proposed an approach using pseudo-relevance feedback of the top retrieved Wikipedia documents to automatically extract the categories of target type.…”

“…M. Koolen [2]'s method and R. Kaptein [4]'s method both have some shortcomings: (1) it need manual participation and a set of topics have same target Wikipedia categories, (2) the filtering model only has two values. R. Kaptein et al [4] proposed an approach using pseudo-relevance feedback of the top retrieved Wikipedia documents to automatically extract the categories of target type. Nevertheless, the number of target Wikipedia categories is insufficient in R. Kaptein [4]'s method which will lower the filtering results.…”

“…R. Kaptein et al [4] proposed an approach using pseudo-relevance feedback of the top retrieved Wikipedia documents to automatically extract the categories of target type. Nevertheless, the number of target Wikipedia categories is insufficient in R. Kaptein [4]'s method which will lower the filtering results.…”