• 유통과학연구
• /
• 제21권3호
• /
• pp.93-100
• /
• 2023

Purpose: The purpose of this paper is to confirm whether stocks belonging to the distribution industry in Korea have reversals, following large daily stock price changes accompanied by large trading volumes. Research design, data, and methodology: We examined whether there were reversals after the event date when large-scale stock price changes appeared for the entire sample of distribution-related companies listed on the Korea Composite Stock Price Index from January 2004 to July 2022. In addition, we reviewed whether the reversals differed depending on abnormal trading volume on the event date. Using multiple regression analysis, we tested whether high trading volume had a significant effect on the cumulative rate of return after the event date. Results: Reversals were confirmed after the stock price shock in the Korean distribution industry and the return after the event date varied depending on the size of the trading volume on the event day. In addition, even after considering both company-specific and event-specific factors, the trading volume on the event day was found to have significant explanatory power on the cumulative rate of return after the event date. Conclusions: Reversals identified in this paper can be used as a useful tool for establishing a trading strategy.

• The Journal of Asian Finance, Economics and Business
• /
• 제8권7호
• /
• pp.383-391
• /
• 2021

This study examines the connectedness between the number of COVID-19 cases in Thailand and trading value among investors in the Stock Exchange of Thailand. Daily data of COVID-19 cases and trading value were sourced from the Thailand ministry of public health and the Stock Exchange of Thailand, from January 12, 2020 to May 11, 2021. This study applies a multiple linear regression analysis to explain the relationship between variables. Empirical evidence clearly shows that the volatility of trading value was affected by COVID-19's new, confirmed, and deaths cases within the first pandemic period more than during the second pandemic period. Nevertheless, during the third pandemic period there is no evidence that the new, confirmed, and deaths cases significantly influenced trading value. Furthermore, the results show that COVID-19's new and deaths cases have a negative coefficient that indicated the trading value-buy/sell decreased in response to COVID-19's new and deaths cases, whereas the confirmed COVID-19 cases have a positive coefficient that indicated the trading value-buy/sell increased in response to COVID's confirmed cases. In summary, this study suggests that the number of COVID-19 cases have a significant impact on the trading value in the short term more than in the intermediate and long term.

• Journal of Information Technology Applications and Management
• /
• 제17권3호
• /
• pp.151-162
• /
• 2010

Recent developments in financial market liberalization and information technology are accelerating the interdependence of national stock markets. This study explores the information spillover effect of the US stock market on the overnight and daytime returns of the Korean stock market. We develop a profitable intra-day trading strategy based on the information spillover effect. Our study provides several important conclusions. First, an information spillover effect still exists from the overnight US stock market to the current Korean stock market. Second, Korean investors overreact to both good and bad news overnight from the US. Therefore, there are significant price reversals in the KOSPI 200 index futures prices from market open to market close. Third, the overreaction effect is different between weekdays and weekends. Finally, the suggested intra-day trading system based on the documented overreaction hypothesis is profitable.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제29권11호
• /
• pp.785-794
• /
• 2002

자동 주식 거래 시스템은 시장 추세의 예측, 투자 종목의 선정, 거래 전략 등 매우 다양한 최적화 문제를 통합적으로 해결할 수 있어야 한다. 그러나 기존의 감독 학습 기법에 기반한 거래 시스템들은 이러한 최적화 요소들의 효과적인 결합에는 큰 비중을 두지 않았으며, 이로 인해 시스템의 궁극적인 성능에 한계를 보인다. 이 논문은 주가의 변동 과정이 마르코프 의사결정 프로세스(MDP: Markov Decision Process)라는 가정 하에, 강화 학습에 기반한 자동 주식 거래 시스템인 R-Trader를 제안한다. 강화 학습은 예측과 거래 전략의 통합적 학습에 적합한 학습 방법이다. R-Trader는 널리 알려진 두 가지 강화 학습 알고리즘인 TB(Temporal-difference)와 Q 알고리즘을 사용하여 종목 선정과 기타 거래 인자의 최적화를 수행한다. 또한 기술 분석에 기반하여 시스템의 입력 속성을 설계하며, 가치도 함수의 근사를 위해 인공 신경망을 사용한다. 한국 주식 시장의 데이타를 사용한 실험을 통해 제안된 시스템이 시장 평균을 초과하는 수익을 달성할 수 있고, 수익률과 위험 관리의 두 가지 측면 모두에서 감독 학습에 기반한 거래 시스템에 비해 우수한 성능 보임을 확인한다.

• The Journal of Asian Finance, Economics and Business
• /
• 제7권9호
• /
• pp.63-74
• /
• 2020

This study investigates the behavior of foreign investors in the Stock Exchange of Thailand (SET) in the time of coronavirus disease 2019 (COVID-19) as to whether trading is abnormal, what strategy is followed, whether herd behavior is present, and whether the actions destabilize the market. Foreign investors' trading behavior is measured by net buying volume divided by market capitalization, whereas the stock market behavior is measured by logged return on the SET index portfolio. The data are daily from Tuesday, August 28, 2018, to Monday, May 18, 2020. The study extends the conditional-regression model in an event-study framework and extracts the unobserved abnormal trading behavior using the Kalman filtering technique. It then applies vector autoregressions and impulse responses to test for the investors' chosen strategy, herd behavior, and market destabilization. The results show that foreign investors' abnormal trading volume is negative and significant. An analysis of the abnormal trading volume with stock returns reveals that foreign investors are not positive-feedback investors, but rather, they self-herd. Although foreign investors' abnormal trading does not destabilize the market, it induces stock-return volatility of a similar size to normal trade. The methodology is new; the findings are useful for researchers, local authorities, and investors.

• 한국정보시스템학회지:정보시스템연구
• /
• 제27권3호
• /
• pp.123-140
• /
• 2018

Purpose The purpose of this study is to explore the optimal trading frequency which is useful for stock price prediction by using deep learning for charting image data. We also want to identify the appropriate time for accurate forecasting of stock price when performing pattern analysis. Design/methodology/approach In order to find the optimal trading frequency patterns and forecast timings, this study is performed as follows. First, stock price data is collected using OpenAPI provided by Daishin Securities, and candle chart images are created by data frequency and forecasting time. Second, the patterns are generated by the charting images and the learning is performed using the CNN. Finally, we find the optimal trading frequency patterns and forecasting timings. Findings According to the experiment results, this study confirmed that when the 10 minute frequency data is judged to be a decline pattern at previous 1 tick, the accuracy of predicting the market frequency pattern at which the market decreasing is 76%, which is determined by the optimal frequency pattern. In addition, we confirmed that forecasting of the sales frequency pattern at previous 1 tick shows higher accuracy than previous 2 tick and 3 tick.

• 경영정보학연구
• /
• 제18권2호
• /
• pp.23-38
• /
• 2016

지금까지 수행된 연구들은 온라인 주식게시판 정보가 주식시장 활동에 미치는 영향의 유무만을 보이는 것에 초점을 맞추었으며, 온라인 주식게시판 정보가 주식투자자에게 매수 의도를 갖도록 하는지, 혹은 매도 의도를 갖도록 하는지에 대해서 연구되지 않았다. 따라서 본 연구의 목적은 온라인 주식게시판 정보가 주로 주식투자자의 어떠한 거래행태를 불러일으키는지 확인하는 것이다. 본 연구의 목적을 달성하기 위하여, 온라인 주식게시판 정보로서 주식 게시물 수를 온라인 구전활동 정도로 보았으며, 매수 및 매도 거래행태로서 주문불균형을 주식투자자의 거래방향성으로 보았다. 그리고 이를 기반으로 온라인 주식게시판의 장내 및 장외 주식게시물 수와 주문불균형 간의 상관관계를 확인하였다. 실증분석을 위하여, KOSPI에 상장된 40개 주식종목에 대한 온라인 주식시판으로부터 3개월 동안의 전체 게시물 46,077개를 수집하였고, 코스콤 데이터베이스로부터 해당 주식 종목에 대한 매수 및 매도 주도거래 데이터를 수집하여 절대 거래횟수 주문불균형 데이터를 설정하였다. 수집한 모든 데이터는 종목 및 시간에 따른 균형 패널데이터(balanced panel data)로 구성하였고, 패널 벡터자기 회귀 분석을 수행하였다. 본 연구의 분석결과를 살펴보면, 온라인 주식게시판의 1, 2일 전(t-1, t-2) 장내 게시물 수는 당일 주문불균형에 양의 영향을 미치는 것으로 나타났다. 그리고 온라인 주식게시판의 1일 전(t-1) 장외 게시물 수는 당일 주문불균형에 양의 영향을 미치는 것으로 나타났다. 즉, 온라인 주식게시판 정보는 주식투자자에게 주로 주식매수 결정에 영향을 미치는 것으로 보여 졌으며, 온라인 주식게시판 정보는 주로 해당 주식을 매수하도록 하는 감성(strong buy or buy sentiment)의 속성을 가진 것으로 추정되었다. 이러한 실증분석 결과를 바탕으로 정보시스템 및 재무행태학 부문의 학술적, 실무적 기여점을 제시한다.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 1999년도 추계학술대회-지능형 정보기술과 미래조직 Information Technology and Future Organization
• /
• pp.337-346
• /
• 1999

In this paper, we propose the rough set approach to extract trading rules able to discriminate between bullish and bearish markets in stock market. The rough set approach is very valuable to extract trading rules. First, it does not make any assumption about the distribution of the data. Second, it not only handles noise well, but also eliminates irrelevant factors. In addition, the rough set approach appropriate for detecting stock market timing because this approach does not generate the signal for trade when the pattern of market is uncertain. The experimental results are encouraging and prove the usefulness of the rough set approach for stock market analysis with respect to profitability.

• 경영과학
• /
• 제30권1호
• /
• pp.43-58
• /
• 2013

Although analysis of charts, which used in stock trading by distinguishing standardized patterns in the movements of stock prices, is simple and easy to use, there can be problems stemming from specific patterns being distinguished as a result of the subjective perspectives of analysts. In accordance with such problems, through the method of template pattern matching, 4 trend reversal patterns were designed and the fitness of the patterns were quantitatively measured. In cases when a stock is purchased when the template pattern fitness value is within a certain range and held for at least 20-days, the average return ratio was analyzed to be higher-with the difference being statistically significant-than the average return ratio attained from trading a stock according to the same method per the Efficient Market Hypothesis. From the results of stock trades of 2 domestic corporations to which the values of the 4 patterns had been applied based on the 4 strategies, it was possible to ascertain differences in the strategy- and pattern-dependent return ratios. Through this study, along with presenting the exceptions for the Efficient Market Hypothesis in stock trading, the fitness level of quantitative chart patterns was measured and the theoretical basis for application of such fitness level was proposed.

• 아태비즈니스연구
• /
• 제11권4호
• /
• pp.317-332
• /
• 2020

Purpose - This study examines the information effect and trading behavior of investors for the 430 stock split data from January 2004 to June 2018 in the Korean stock market. Design/methodology/approach - The stock split samples are classified into two groups by split ratio as well as three groups by price level prior to split. We also investigate the trading behavior of investors categorized by institutional versus individual investors. Findings - First, we find a significantly positive information effect on the announcement day. In particular, the information effect is more distinct in the group of larger split ratio and higher price level of stocks. Second, we find a huge increase in turnover following the stock splits, which mainly results from the trading by individual investors. Also, the increase in turnover by individual investors is evident in the group of larger split ratio and higher price level of stocks. Third, the stock splits have a negative impact on the long-term stock performance. The negative buy-and-hold abnormal return(BHAR) makes no difference in the groups by split ratio as well as price level of stocks. Lastly, we find individual investors tend to buy splitted stocks, which exhibit the long-term under-performance. Research implications or Originality - The results in this paper suggest that the liquidity hypothesis is not supported in the Korean stock splits. In addition, we observe that individual investors are exposed to losses due to their unfavorable trading behavior following the stock split.