A Competency Mining Method Based on Latent Dirichlet Allocation (LDA) Model. (November 2020)

• Record Type: Journal Article
• Authors: Wu, Jiying; Son, Guandong; Wang, Sihui
• Is Part Of: Journal of physics. Volume 1682(2020)
• Publication Date: 2020-11
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1682/1/012059
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: A text mining approach based on latent Dirichlet allocation (LDA) is proposed to analyze the competency characteristics. First, we briefly introduce the principle and hypothesis of latent Dirichlet allocation (LDA) model. Second, we elaborate the idea of using LDA topic model to extract competency, Then, we use Chinese text materials such as biographies and interviews of Chinese scientific and technological talents as data sources to conduct experiments, and obtain four competency topics including knowledge, attitude, quality and values. The research results preliminarily verify the effectiveness of latent Dirichlet allocation (LDA) model in competency research, but there are still many details to be improved in the future.