TopicSifter: Interactive Search Space Reduction Through Targeted Topic Modeling
This addresses the need for efficient and accurate document retrieval in domains where missing relevant documents is critical, but it appears incremental as it builds on existing topic modeling methods.
The paper tackles the problem of large-scale document retrieval where users need to focus on specific aspects rather than entire corpora, and presents TopicSifter, a visual analytics system that uses targeted topic modeling with relevance feedback to achieve high recall, though no concrete numbers are provided.
Topic modeling is commonly used to analyze and understand large document collections. However, in practice, users want to focus on specific aspects or "targets" rather than the entire corpus. For example, given a large collection of documents, users may want only a smaller subset which more closely aligns with their interests, tasks, and domains. In particular, our paper focuses on large-scale document retrieval with high recall where any missed relevant documents can be critical. A simple keyword matching search is generally not effective nor efficient as 1) it is difficult to find a list of keyword queries that can cover the documents of interest before exploring the dataset, 2) some documents may not contain the exact keywords of interest but may still be highly relevant, and 3) some words have multiple meanings, which would result in irrelevant documents included in the retrieved subset. In this paper, we present TopicSifter, a visual analytics system for interactive search space reduction. Our system utilizes targeted topic modeling based on nonnegative matrix factorization and allows users to give relevance feedback in order to refine their target and guide the topic modeling to the most relevant results.