Year: 2021
Type: article
Abstract: Document summarization plays a vital role in Natural Language Processing (NLP), which is used to condense extensive textual content into concise versions while preserving the essential information. This paper offers a comprehensive survey of unsupervised document summarization techniques, with a particular emphasis on graph-based methods. We explore the foundational principles of graph-based summarization and illustrate how documents are represented as nodes in a graph which is utilized for extraction of crucial sentences based on their centrality and connectivity within the graph structure. Furthermore, we also conduct a comparative analysis between graph-based methods and traditional approaches like TF-IDF and Latent Dirichlet Allocation (LDA) to identify their respective strengths and limitations. Through this exploration, we aim to provide insights into the efficacy of graph-based techniques in document summarization and guide future research endeavors in this domain.
Author: Akshata Upadhye
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4403565682