moving average

The present research proposes an automatic system based on moving average (MA) and fuzzy logic to recognize technical analysis patterns including head and shoulder patterns, triangle patterns and broadening patterns in the Tehran Stock Exchange. The automatic system was used on 38 indicators of Tehran Stock Exchange within the period 2014-2017 in order to evaluate the effectiveness of technical patterns. Having compared the conditional distribution of daily returns under the condition of the discovered patterns and the unconditional distribution of returns at various levels of confidence driven from fuzzy logic with the mean returns of all normalized market indicators, we observed that in the desired period, after recognizing the pattern, all patterns investigated at the confidence level 0.95 with a fuzzy point 0.5 contained useful information, practically leading to abnormal returns.

Traders in the capital market always seek methods to make full use of available information and combine them to find the best buying and selling strategy. The present study uses a linear hybrid system to combine 106 signals from moving averages oscillators and RSI signals in the technical analysis along with two buy and sell bonds. In addition, the system has correctional property and modifies its parameters over time and according to new information. The result of the research on the Tehran Exchange overall index in the period 1380 to 1397 indicates that the system after the optimal training on training data has an average of daily returns of 0/0025, 0/0048 risk, and a daily Sharp ratio of 0/52, which is better than the individual performance of each signal and market performance in daily average return and sharp ratio criterion.

Designing trading systems with good returns is critical for capital market investors. Trading systems are often based on a combination of several tools to use their combined information. For the first time in Iran, the present study aimed to propose a pattern detection algorithm for a flag pattern based on Japanese candlestick charts and their arrangement. By recognizing the pattern and if the 4- and 10-day moving average is confirmed, a shopping position is developed, and the selling time is determined based on an optimized and dynamic process commensurate with price changes and the data scale. Our objective was to address the question of whether the returns resulting from this strategy have a more significant positive return compared to the purchase and maintenance strategy. The research sample included the daily information of 16 active companies of basic metals in Tehran Stock Exchange during 2007-2019, extracted from the database of Novin Rahavard software. Data analysis was performed in MATLAB software, and the obtained experimental evidence was described using t-test. According to the results, the research strategy had a higher performance in terms of returns and risks compared to the market.