کشف قانون تجارت بازار سهام با استفاده از تشخیص الگو و تجزیه و تحلیل فنی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|28379||2007||12 صفحه PDF||سفارش دهید||8860 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Expert Systems with Applications, Volume 33, Issue 2, August 2007, Pages 304–315
This study examines the potential profit of bull flag technical trading rules using a template matching technique based on pattern recognition for the Nasdaq Composite Index (NASDAQ) and Taiwan Weighted Index (TWI). To minimize measurement error due to data snooping, this study performed a series of experiments to test the effectiveness of the proposed method. The empirical results indicated that all of the technical trading rules correctly predict the direction of changes in the NASDAQ and TWI. This finding may provide investors with important information on asset allocation. Moreover, better bull flag template price fit is associated with higher average return. The empirical results demonstrated that the average return of trading rules conditioned on bull flag significantly better than buying every day for the study period, especially for TWI.
Developing a model for predicting returns is an important goal for academics and practitioners. Fundamental and technical analysis has longed aimed to devise trading rules suitable for application on stock markets. A significant body of literature exists on fundamental and technical analysis in various financial domains. Results obtained in the 1960s and 1970s supported the “Efficient Market Hypothesis”, which states that the efficient nature of financial markets should mean that market data does not contain any discernable and exploitable patterns (Alexander, 1964, Fama and Blume, 1966 and Jensen and Bennington, 1970). Therefore, impulses from new information cannot be predicted. The market prices are best described as a random walk, and past price and volume information are worthless for predicting future market price behavior. However, some recent results since the 1980s have appeared to indicate otherwise. Well-known anomalies involve abnormal returns associated with: unexpected earnings announcements, firm size, the month of January, the day of the week, and so on. Additionally, the behavior finance literature uses a conservative bias and investor overconfidence to explain evidence of market underreaction or overreaction to information documented by DeBondt and Thaler, 1985 and Jegadeesh and Titman, 1993, among others. The studies of abnormal return and behavior finance indicated that historical prices can help in predicting future prices. Sharpe, Alexander, and Bailey (1995) summarized some observations regarding the recent evidence in technical analysis, stating “the apparent success of these (technical) strategies offers a challenge to those who contend that the stock market is highly efficient”. Consequently, numerous financial researches have progressively employed a positive and careful attitude to probe into technical analysis. A fairly comprehensive literature related to technical analysis in various financial domains has addressed numerous effective evidences that trading success can be achieved with technical analysis. Technical analysis studies records or charts of past stock prices, hoping to identify patterns that can be exploited to achieve excess profits. Academic study of technical analysis has mainly adopted quantitative indicators as prediction variables, for example relative strength index, moving average and so on. Meanwhile, charting pattern, for example head-and-shoulder, flag, etc. are comparatively rare. Lo, Mamaysky, and Wang (2000) considered many technical quantitative indicators that may find it easier to detect algorithmically-moving average, support and resistance levels, oscillators, and so on, but that those charting patterns are most difficult to quantify analytically. Nevertheless, complying with the development of computer technology and cross-domain research, academic study has gradually paid increasing attention to pattern analysis for investment decision, including Lo et al. (2000) testing price charting patterns using kernel regression for the identification of ten patterns. Leigh, Purvis et al., 2002, Leigh, Modani et al., 2002 and Leigh et al., 2004 implemented a variation of the bull flag stock chart using a template matching technique based on pattern recognition. All of these researches showed that trading success can be achieved with charting patterns. This study developed a new template grid, bull flag, and a method of calculating fit value using a template matching technique from pattern recognition. This study concentrates on identifying increasing price value, regardless of the nature of the preceding or accompanying news, using a version of the bull flag charting pattern. The detection of this bull flag pattern in the time series of index values for the Nasdaq Composite Index (hereafter NASDAQ) and Taiwan Weighted Index (hereafter TWI) becomes a buy signal. This study fills a gap in the literature, since no previous study has applied the bull flag trading rules to the Taiwanese market, which is an emerging market, and the NASDAQ, which is a developed market. This study also improved the methodology applied by previous studies on this area.1 For empirical analysis, to minimize measurement error due to data snooping, this study applies the method of Brock, Lakonishok, and LeBaron (1992) to use a long data series for NASDAQ and TWI, and reports results for various fitting windows, holding horizons, and threshold values. Moreover, this study tests performance consistency for various non-overlapped sub-periods. The empirical results indicated that all of the technical trading rules correctly predict the direction of changes in the index series. These findings may provide investors with important asset allocation information. Additionally, the buy signal with better bull flag template price fit is associated with higher average returns. The empirical results demonstrated that trading rules based on bull flag (conditional trading rules) significantly better than buying every day (unconditional trading rules) of the study period, especially for TWI. The remainder of this paper is organized as follows. Section 2 reviews the previous literature on technical trading rules. The proposed method is then described in Section 3. Next, Section 4 describes the data and results of the empirical investigation. Finally, Section 5 offers concluding remarks.
نتیجه گیری انگلیسی
This study examined the potential profit by applying technical trading rules, bull flag, using template matching technique based on pattern recognition in the NASDAQ and TWI. To minimize measurement error due to data snooping, this study designed a series of experiments for testing the effectiveness of the proposed method. The empirical results indicated that all of the bull flag trading rules, regardless of their statistical significance, correctly predict the direction of changes in the index series of the NASDAQ and TWI. This finding may provide investors with important information on asset allocation. Furthermore, given shorter holding periods, bull flag trading rules conditioned on the shorter fitting window or better bull flag template price fit generate higher excess profit. The empirical results also demonstrate that bull flag trading rules conditioned on higher threshold value generate higher excess profit, but with fewer trading days. Owing to the practicality of the trading procedure, this study suggests setting the threshold value to 2–3. Overall, the technical trading rules have greater forecasting power for Taiwanese stock markets, where excess profits are more significant across experiments, than for more developed markets such as the NASDAQ, where excess profits are positive but less significant. These results are consistent with the results reported by Bessembinder and Chan, 1995 and Ratner and Leal, 1999, the technical trading rules, moving average, can be profitable in some Asian countries, such as Taiwan and Thailand. Notably, our study used charting patterns to test the effect, rather than the quantitative indicators used in previous studies.