Community Detection과 Clustering을 이용한 주식 시장 군집화Comparisons of Community Detection and Clustering in Stock Market Classifications
- Other Titles
- Comparisons of Community Detection and Clustering in Stock Market Classifications
- Authors
- 박건빈; 박성훈; 박민우; 이정수; 윤지환; 한성원
- Issue Date
- 2020
- Publisher
- 대한산업공학회
- Keywords
- KOSPI Market Classification; Community Detection; Clustering; Silhouettes
- Citation
- 대한산업공학회지, v.46, no.6, pp.626 - 636
- Indexed
- KCI
- Journal Title
- 대한산업공학회지
- Volume
- 46
- Number
- 6
- Start Page
- 626
- End Page
- 636
- URI
- https://scholar.korea.ac.kr/handle/2021.sw.korea/59149
- DOI
- 10.7232/JKIIE.2020.46.6.626
- ISSN
- 1225-0988
- Abstract
- When investing in companies or deciding a portfolio, investors often consider the company’s industry or theme, especially Industry Classification in KOSPI Industry Group Indices. However, the classification is not based on statistical analysis using the stock price of companies. Therefore, the current study aims to suggest a new market classification based on network analysis to diversify portfolios. In the study, we applied a lasso and adaptive lasso in KOSPI data (Jan 2012~Jul 2019) to estimate an undirected graph. Then, We applied community detection and clustering analysis to classify companies in the estimated graph. As a result, most clusters include companies with different sectors in KOSPI Industry Group Indices which are closely related in reality. Moreover, based on silhouettes, new market classification results outperform the existing classification. Thus, new market classification can be a better alternative to the existing classification.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - Graduate School > Graduate School of management of technology > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.