Proximity-based document representation for named entity retrieval

Abstract

One aspect in which retrieving named entities is different from retrieving documents is that the items to be retrieved -persons, locations, organizations -are only indirectly described by documents throughout the collection. Much work has been dedicated to finding references to named entities, in particular to the problems of named entity extraction and disambiguation. However, just as important for retrieval performance is how these snippets of text are combined to build named entity representations.We focus on the TREC expert search task where the goal is to identify people who are knowledgeable on a specific topic. Existing language modeling techniques for expert finding assume that terms and person entities are conditionally independent given a document. We present theoretical and experimental evidence that this simplifying assumption ignores information on how named entities relate to document content. To address this issue, we propose a new document representation which emphasizes text in proximity to entities and thus incorporates sequential information implicit in text. Our experiments demonstrate that the proposed model significantly improves retrieval performance. The main contribution of this work is an effective formal method for explicitly modeling the dependency between the named entities and terms which appear in a document.

A User-Oriented Model for Expert Finding

“…Building on either candidate or document models, further refinements to estimating the association of a candidate with the topic of expertise have been explored. For example, instead of capturing the associations at the document level, they may be estimated at the paragraph or snippet level [7,20,24]. Other extensions incorporate additional forms of evidence through the use of priors [14], document structure [33], hierarchical, organizational, and topical context and structure [6,23], and Web data [25].…”

Semantic annotation based exploratory search for information analysts

“…As a semantic category, named entities (NEs) act as pointers to real world entities such as locations, organizations, people, or events (Petkova & Croft, 2007). Because NEs can provide much richer semantic content than most vocabulary words, they have been studied extensively in various language processing and information access tasks.…”

Search for expertise : going beyond direct evidence

“…In another approach, the partial relevance of each query term instance found in a document contributed to the probability of a candidate's expertness, but proportionally to its word distance from the nearest mention of the candidate in this document. Different distance functions have been applied and some of them lead to window-based approaches (Petkova and Croft, 2007).…”

“…While one approach measures the degree of proximity of a term to the query terms in the scope of a document (Gao et al, 2002), the other also takes the sequential order of terms into account (Metzler and Croft, 2007). Once, it was shown for expert finding that the overall pairwise distance between a candidate's mention and the query terms in a document expresses the degree of association between the document and the candidate (Petkova and Croft, 2007). At the same time, the importance of the order in which personal identifiers and query terms occur in documents was never studied to the best of our knowledge.…”

“…Furthermore, methods estimating relevance of the textual content related to a person on the lower and hence less ambiguous level (e.g. paragraph or sentence level) usually more effective (Petkova and Croft, 2007;) (see Section 2.1.2). Most of these approaches ignore the complexity of link structure among persons and documents and hence do not consider the expertness of directly and indirectly linked persons as well as the relevance of documents found not in the immediate neighborhood of a candidate.…”