Relevance models for topic detection and tracking

Abstract

We extend relevance modeling to the link detection task of Topic Detection and Tracking (TDT) and show that it substantially improves performance. Relevance modeling, a statistical language modeling technique related to query expansion, is used to enhance the topic model estimate associated with a news story, boosting the probability of words that are associated with the story even when they do not appear in the story. To apply relevance modeling to TDT, it had to be extended to work with stories rather than short queries, and the similarity comparison had to be changed to a modified form of Kullback-Leibler. We demonstrate that relevance models result in very substantial improvements over the language modeling baseline. We also show how the use of relevance modeling makes it possible to choose a single parameter for within-and cross-mode comparisons of stories.

Story Link Detection in Turkish Corpus

“…In order to identify the effects of the number of index terms on the match performance, tests were repeated for 1, 2, 3, 4,5,10,15,20,25,30,35,40,45,50,55,60,65,70,75,80,85,90,95,100,125,150,175,200,225,250,275, 300, 400, 500 and 1000 terms respectively. In addition, logical operators AND, and OR were applied on the results obtained through the use of VSM and RM methods so that the effects of Boolean operators AND and OR on precision and recall measures can be seen.…”

“…• Story Link Detection; to distinguish if the two different stories are on the same subject or not. In the TDT studies, the story link detection task is reported to have a critical role [5][6] [9]. Carrying out the story link detection task successfully is expected to solve many problems in TDT [10].…”

“…TDT algorithms based on IR models detect new unreported stories and organize them temporally, link incoming news items with previously detected stories on the same topic [3], and monitor news streams online till stories peter out [4]. Story link detection (SLD) algorithms play a key role in establishing linkages between stories discussing the same subjects [5] [6]. Detected stories along with their updates can be delivered to the users based on their current interests and past behaviors.…”

A Performance Prediction Approach to Enhance Collaborative Filtering Performance

“…Because of that, the clarity score method has been widely used for query performance prediction in the area. Some applications include query expansion (anticipating poorly performing queries as good candidates to be expanded), rank fusion, link extraction in topic detection and tracking [15], and document segmentation [8]. A prolific sequel of variants and enhancements on the notion of clarity followed the original works [8,11].…”

Predicting Neighbor Goodness in Collaborative Filtering

“…Because of that, the clarity score method has been widely used for query performance prediction in the area. Some applications include query expansion (anticipating poorly performing queries as good candidates to be expanded), rank aggregation, link extraction in topic detection and tracking [16], and document segmentation [8]. A prolific sequel of variants and enhancements on the notion of clarity followed the original works [8,11].…”