2013The Journal of the Korea Contents AssociationOpen access

Analysis of the Stock Market Network for Portfolio Recommendation

Yunjung Lee, Gyun Woo

Open full text 5 citations

Abstract

주식시장은 시간에 따라 계속 변하고 특별한 이유 없이 주가가 급등하거나 급락하는 사건들이 발생하기도 한다. 이런 이유로 주식시장은 복잡계로 인식되고 있으며 주가 변동을 예측하는 것은 어려운 일이다. 이 논문에서는 주식시장을 개별 주식들의 네트워크로 이해하고 시간에 따라 변하는 한국 주식시장 네트워크를 분석하였다. 코스피200 지수를 구성하는 137개 회사의 주식들을 대상으로 주식 사이의 상관관계를 측정한 결과 주식 간 상관관계가 매우 높을 때 주가가 급락하는 경향이 있는 것으로 나타났다. 또한, 우리는 이러한 네트워크 분석 결과를 바탕으로 주식 포트폴리오를 구성하는 방법을 제안한다. 제안 방법으로 구성된 포트폴리오의 효율성을 보이기 위해 실제 주식들을 대상으로 모의 투자 실험을 수행하였고, 마코위츠의 효율적 포트폴리오 구성 알고리즘을 이용해 구성한 포트폴리오의 수익률과 비교하였다. 실험 결과 제안 방법으로 구성된 포트폴리오는 평균적으로 약 10.6%의 수익률을 보였으며, 같은 기간 마코위츠의 효율적 포트폴리오의 수익률보다 약 3.7% 높으며, 코스피200 수익률보다 약 5.6% 정도 높게 나타났다. The stock market is constantly changing and sometimes a slump or a sudden rising in stocks happens without any special reason. So the stock market is recognized as a complex system and it is hard to predict the change on stock prices. In this paper we consider the stock market to a network consisting of stocks. We analyzed the dynamics of the Korean stock market network and evaluated the changing of the correlation between shares consisting of the time series data of 137 companies belong to KOSPI200. Our analysis shows that the stock prices tend to plummet when the correlation between stocks is very high. We propose a method for recommending the stock portfolio based on the analysis of the stock market network. To show the effectiveness of the recommended portfolio, we conducted the simulated stock investment and compared the recommended portfolio with the efficient portfolio proposed Markowitz. According to the experiment results, the rate of return of the portfolio is about 10.6% which is about 3.7% and 5.6% higher than the average rate of return of the efficient portfolio and KOSPI200 respectively.

Open-access reader

About this research paper

What this paper is about

주식시장은 시간에 따라 계속 변하고 특별한 이유 없이 주가가 급등하거나 급락하는 사건들이 발생하기도 한다. 이런 이유로 주식시장은 복잡계로 인식되고 있으며 주가 변동을 예측하는 것은 어려운 일이다. 이 논문에서는 주식시장을 개별 주식들의 네트워크로 이해하고 시간에 따라 변하는 한국 주식시장 네트워크를 분석하였다. 코스피200 지수를 구성하는 137개 회사의 주식들을 대상으로 주식 사이의 상관관계를 측정한 결과 주식 간 상관관계가 매우 높을 때 주가가 급락하는 경향이 있는 것으로 나타났다. 또한, 우리는 이러한 네트워크 분석 결과를 바탕으로 주식 포트폴리오를 구성하는 방법을 제안한다. 제안 방법으로 구성된 포트폴리오의 효율성을 보이기 위해 실제 주식들을 대상으로 모의 투자 실험을 수행하였고, 마코위츠의 효율적 포트폴리오 구성 알고리즘을 이용해 구성한 포트폴리오의 수익률과 비교하였다. 실험 결과 제안 방법으로 구성된 포트폴리오는 평균적으로 약 10.6%의 수익률을 보였으며, 같은 기간 마코위츠의 효율적 포트폴리오의 수익률보다 약 3.7% 높으며, 코스피200 수익률보다 약 5.6% 정도 높게 나타났다. The stock market is constantly changing and sometimes a slump or a sudden rising in stocks happens without any special reason. So the stock market is recognized as a complex system and it is hard to predict the change on stock prices. In this paper we consider the stock market to a network consisting of stocks. We analyzed the dynamics of the Korean stock market network and evaluated the changing of the correlation between shares consisting of the time series data of 137 companies belong to KOSPI200. Our analysis shows that the stock prices tend to plummet when the correlation between stocks is very high. We propose a method for recommending the stock portfolio based on the analysis of the stock market network. To show the effectiveness of the recommended portfolio, we conducted the simulated stock investment and compared the recommended portfolio with the efficient portfolio proposed Markowitz. According to the experiment results, the rate of return of the portfolio is about 10.6% which is about 3.7% and 5.6% higher than the average rate of return of the efficient portfolio and KOSPI200 respectively.

Why it matters

OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

주식시장은 시간에 따라 계속 변하고 특별한 이유 없이 주가가 급등하거나 급락하는 사건들이 발생하기도 한다. 이런 이유로 주식시장은 복잡계로 인식되고 있으며 주가 변동을 예측하는 것은 어려운 일이다. 이 논문에서는 주식시장을 개별 주식들의 네트워크로 이해하고 시간에 따라 변하는 한국 주식시장 네트워크를 분석하였다. 코스피200 지수를 구성하는 137개 회사의 주식들을 대상으로 주식 사이의 상관관계를 측정한 결과 주식 간 상관관계가 매우 높을 때 주가가 급락하는 경향이 있는 것으로 나타났다. 또한, 우리는 이러한 네트워크 분석 결과를 바탕으로 주식 포트폴리오를 구성하는 방법을 제안한다. 제안 방법으로 구성된 포트폴리오의 효율성을 보이기 위해 실제 주식들을 대상으로 모의 투자 실험을 수행하였고, 마코위츠의 효율적 포트폴리오 구성 알고리즘을 이용해 구성한 포트폴리오의 수익률과 비교하였다. 실험 결과 제안 방법으로 구성된 포트폴리오는 평균적으로 약 10.6%의 수익률을 보였으며, 같은 기간 마코위츠의 효율적 포트폴리오의 수익률보다 약 3.7% 높으며, 코스피200 수익률보다 약 5.6% 정도 높게 나타났다. The stock market is constantly changing and sometimes a slump or a sudden rising in stocks happens without any special reason. So the stock market is recognized as a complex system and it is hard to predict the change on stock prices. In this paper we consider the stock market to a network consisting of stocks. We analyzed the dynamics of the Korean stock market network and evaluated the changing of the correlation between shares consisting of the time series data of 137 companies belong to KOSPI200. Our analysis shows that the stock prices tend to plummet when the correlation between stocks is very high. We propose a method for recommending the stock portfolio based on the analysis of the stock market network. To show the effectiveness of the recommended portfolio, we conducted the simulated stock investment and compared the recommended portfolio with the efficient portfolio proposed Markowitz. According to the experiment results, the rate of return of the portfolio is about 10.6% which is about 3.7% and 5.6% higher than the average rate of return of the efficient portfolio and KOSPI200 respectively.

Key concepts: Stock market, Portfolio, Stock (firearms), Financial economics, Econometrics, Portfolio insurance, Stock market bubble, Business

Related papers

Back to paper searchBrowse research topicsOriginal source
Analysis of the Stock Market Network for Portfolio Recommendation — Research Paper | ScholarLens