Finding answers to COVID-19-specific questions: An information retrieval system based on latent keywords and adapted TF-IDF.

The scientific community has reacted to the COVID-19 outbreak by producing a high number of literary works that are helping us to understand a variety of topics related to the pandemic from different perspectives. Dealing with this large amount of information can be challenging, especially when researchers need to find answers to complex questions about specific topics. We present an Information Retrieval System that uses latent information to select relevant works related to specific concepts. By applying Latent Dirichlet Allocation (LDA) models to documents, we can identify key concepts related to a specific query and a corpus. Our method is iterative in that, from an initial input query defined by the user, the original query is expanded for each subsequent iteration. In addition, our method is able to work with a limited amount of information per article. We have tested the performance of our proposal using human validation and two evaluation strategies, achieving good results in both of them. Concerning the first strategy, we performed two surveys to determine the performance of our model. For all the categories that were studied, precision was always greater than 0.6, while accuracy was always greater than 0.8. The second strategy also showed good results, achieving a precision of 1.0 for one category and scoring over 0.7 points overall. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Information Science is the property of Sage Publications, Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)