Prediction-Based Portfolio Optimization Model for Iran’s Oil Dependent Stocks Using Data Mining Methods

Abstract

This study applied a prediction-based portfolio optimization model to explore the results of portfolio predicament in the Tehran Stock Exchange. To this aim, first, the data mining approach was used to predict the petroleum products and chemical industry using clustering stock market data. Then, some effective factors, such as crude oil price, exchange rate, global interest rate, gold price, and S&P 500 index, were used to estimate each industry index using Radial Basis Function and Multi-Layer Perceptron neural networks. Finally, by comparing the validation ratios in a bullish market using K-Means, SOM, and Fuzzy C-means clustering algorithms, the best algorithm was employed to predict indicators for each industry. The sample was collected between December 15, 2008, and April 25, 2018. The results revealed that the Multi-Layer Perceptron algorithm had the highest accuracy and was the best option for portfolio predicament. However, the Fuzzy C-means algorithm produced the best clusters. Practical results showed that Sepahan oil and Kharg petrochemical stocks were the most important stocks in the short term while Kharg petrochemical, Fannavaran petrochemical, and Tehran oil refinery stocks made higher contributions in a stock portfolio in the medium- or long-term.