A Suggestion on the LDA-Based Topic Modeling Technique Based on ElasticSearch for Indexing Academic Research Results

Abstract

Most academic researchers use the academic information system when they want to write a reference, such as a related research for a paper. Specific classification rules are applied based on vast amounts of data and the latest references to classify and search keywords. Meta information is designed for specific classification rules and search results are restructured. The search results can be classified and rearranged to suit academic research paper keywords by applying the restructured classification system and the LDA-based topic modeling technique. To implement this, the ElasticSearch classification method and topic-based LDA model were applied to extract the characteristics of academic papers in this study. Stable topics that could detect topic estimation and keyword search results within the minimum time were extracted to classify the paper search results. In addition, by analyzing the distribution of document weight among topics, the system performance was proven to be excellent.

Keywords

• machine learning
• natural language
• text-mining
• LDA model
• ElasticSearch
• meta-data