Enhancing Temporal Information Retrieval through Contextual Query Reformulation
Vishal Gupta, Ashutosh Dixit, Shilpa Sethi · 2023
As the amount of data is rising exponentially on the World Wide Web, it becomes challenging to retrieve information that is relevant to a specific period. In this paper, a query reformulation architecture is proposed to extract and rank web pages based on the temporal context of the query. If a temporal query is given, the proposed architecture extracts the web pages as well as trending hashtags and tweets from X related to the query. After preprocessing of extracted web pages and trending hashtags as well as tweets, contextual terms are extracted from them and added to the original query to form a ‘Reformulated query’. The reformulated query is then used to re-rank the extracted web pages based on the similarity score between the reformulated query and extracted web pages. The proposed architecture is evaluated on the basis of MAP, MRR, and nDCG and it has been observed that the proposed architecture outperforms the baseline models in terms of relevance and trend scores.