Inferring query aspects from reformulations using clustering

Dang, Van; Xue, Xiaobing; Croft, W. Bruce

doi:10.1145/2063576.2063904

Cited by 25 publications

(24 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Radlinski et al [33] infer query aspects from query logs, using clicks and session information to model the relations between queries. Dang et al [20] infer query intent from anchor text and Web ngrams. In both studies, clustering has been applied to the extracted aspects or intents, so that topically redundant or similar entities (queries, anchor text, or Web ngrams) are grouped together.…”

Section: Related Workmentioning

confidence: 99%

“…[19]. We use the anchor texts from the ClueWeb09 collection [27] to construct an anchor text graph G A , using the method described by Dang et al [20]. As preprocessing, we remove all anchors that are connected to only one URL.…”

Section: Anchor Textsmentioning

confidence: 99%

“…This paper focuses on the discovery of subtopics, also referred to as query intent, facets, and sub-queries in the literature. Subtopics have been extracted internally, e.g., from documents retrieved in response to the query [7,8,26,29], and externally, e.g., from Web resources such as query logs [1,20,21,33,35]. Internally modelled subtopics are often implicitly represented, e.g., as latent topics or document clusters, while externally modelled subtopics are often explicitly represented, e.g., as reformulated queries 1 .…”

Section: Introductionmentioning

confidence: 99%

“…In particular, there has been much work that exploits explicit subtopics using various types of Web resources. Some take suggested queries from commercial search engines as subtopics of a query [35]; others mine the subtopics from resources such as taxonomies [1,22], query logs [21,33], anchor text and Web ngrams [20].…”

Section: Introductionmentioning

confidence: 99%

“…For instance, all subtopics discovered in Web resources, e.g., all reformulations of a query, are treated as independent subtopics, regardless the fact that some may be synonymous [21,35]. In some studies additional steps such as clustering were taken in order to avoid redundancy and find better defined topics [20,33]. On the other hand, implicitly represented subtopics such as topic models or clusters constructed from document content convey much information about the relations or the structure of the topics present in the documents.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Combining implicit and explicit topic representations for result diversification

Hollink

Vries

2012

Proceedings of the 35th International ACM SIGIR Conference on Research and Development in Information Retrieval

View full text Add to dashboard Cite

Result diversification deals with ambiguous or multi-faceted queries by providing documents that cover as many subtopics of a query as possible. Various approaches to subtopic modeling have been proposed. Subtopics have been extracted internally, e.g., from retrieved documents, and externally, e.g., from Web resources such as query logs. Internally modeled subtopics are often implicitly represented, e.g., as latent topics, while externally modeled subtopics are often explicitly represented, e.g., as reformulated queries.We propose a framework that: i) combines both implicitly and explicitly represented subtopics; and ii) allows flexible combination of multiple external resources in a transparent and unified manner. Specifically, we use a random walk based approach to estimate the similarities of the explicit subtopics mined from a number of heterogeneous resources: click logs, anchor text, and web n-grams. We then use these similarities to regularize the latent topics extracted from the top-ranked documents, i.e., the internal (implicit) subtopics. Empirical results show that regularization with explicit subtopics extracted from the right resource leads to improved diversification results, indicating that the proposed regularization with (explicit) external resources forms better (implicit) topic models. Click logs and anchor text are shown to be more effective resources than web n-grams under current experimental settings. Combining resources does not always lead to better results, but achieves a robust performance. This robustness is important for two reasons: it cannot be predicted which resources will be most effective for a given query, and it is not yet known how to reliably determine the optimal model parameters for building implicit topic models.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Anchor Textsmentioning

confidence: 99%