Optimizing stock portfolios by comparing different technical patterns

In recent years, different studies, using heuristic approaches for portfolio optimization, have been done to maximize the return and minimize the risk. In this study, we employed nine useful indicators, including SMA, EMA, ROC, OBV, RSI, MACD, TSI, HMA, Fibonacci Retracement, and genetic algorithm (GA), to construct an expert system that optimizes the portfolio automatically. In this system, Buy, Sell or Hold signals are produced based on the weighted combination of mentioned indicators and under the specified thresholds for each one. After that, signals (Buy/Sell/Hold) feed to the GA to optimize portfolio return/risk by changing indicator weights. This algorithm repeats continuously to optimize indicators weights for system and qualified stocks are selected according to optimized weights. Outputs of this expert system have been compared with Tehran Stock Exchange Index (Value-weighted and Equal-weighted Indices) from March 2013 until June 2021.The empirical results show that this expert system outperforms the Buy-and-Hold strategy (TSE indices) in terms of risk and return.