Portfolio Selection by Optimizing Risk and Return Based on Complex Network Analysis (Case Study: Tehran Stock Exchange)

Document Type : Original Article


Amirabad, North Kargar Street, Faculty of New Sciences and Technologies, University of Tehran



Nowadays, complex networks are applied for analyzing a huge body of data. Since the stock market has large data that are constantly fluctuating, it is highly difficult to analyze these data and manage the stock purchase and sale for investors. In this research, complex network is applied to select stock portfolios in order to facilitate market analysis and decision making in business relationships, and reduce the risk of inaccurate decisions. For this purpose, Tehran Stock Exchange was selected and subsequently, the latest data were collected over six consecutive years. Afterwards, a stock return correlation network was developed. According to the community detection, cohort groups were identified and then, a stock was selected from each community by designing an optimization model from the network centralities, risk and returns. Finally, for checking the accuracy of the selected portfolio, the portfolio performance in two ways with and without risk was compared with the performance of the TEPIX index. Results of this study showed that complex networks played a very effective role in selecting stock portfolios with high returns and low risk by visualizing lots of stocks in one network picture and facilitate global characteristics analysis across the network.