Addressing topic modeling with a multi-objective optimization approach based on swarm intelligence

作     者：Gonzalez-Santos, Carlos Vega-Rodriguez, Miguel A. Perez, Carlos J. 

作者机构：Univ Extremadura Dept Math Campus Univ S-N Caceres 10003 Spain Univ Extremadura Dept Comp & Commun Technol Campus Univ S-N Caceres 10003 Spain 

出 版 物：《KNOWLEDGE-BASED SYSTEMS》 (知识库系统)

年 卷 期：2021年第225卷

页      面：107113-107113页

核心收录：

学科分类：08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Ministry of Science, Innovation and Universities - Spain [PID2019-107299GB-I00/AEI/10.13039/501100011033, MTM2017-86875-C3-2-R] State Research Agency - Spain [PID2019-107299GB-I00/AEI/10.13039/501100011033, MTM2017-86875-C3-2-R] Junta de Extremadura -Spain [GR18090, GR18108] European Union (European Regional Development Fund) 

主　　题：Artificial bee colony Evolutionary computing Latent Dirichlet allocation Multi-objective optimization Text analysis Topic modeling 

摘      要：Topic modeling is a growing field within the area of text analysis that extracts underlying topics from document collections. Several objectives can be simultaneously considered when designing an approach for topic modeling. A multi-objective optimization approach based on the swarm intelligence of a bee colony (MOABC, Multi-Objective Artificial Bee Colony) has been designed, implemented, and tested. This new approach has been evaluated by using documents from the Reuters-21578 and TagMyNews datasets. Three objective functions (coherence, coverage, and perplexity) and three multi objective metrics (hypervolume, set coverage, and distance to the ideal point) have been considered in two topic scenarios. Results show that MOABC provides relevant improvements with respect to LDA (Latent Dirichlet Allocation, the most used approach for topic modeling) and MOEA/D (Multi-Objective Evolutionary Algorithm based on Decomposition, the only multi-objective approach published to date). This demonstrates that the multi-criteria nature of topic modeling should be exploited with multi-objective optimization approaches. (C) 2021 Elsevier B.V. All rights reserved.Y