۱- Introduction

۲- Methodology

۳- Results

۴- Discussion and conclusion

References

Abstract

A big data set can often be illustrated by the nodes and edges of a big network. A large volume of data is generally produced by the stock market, and complex networks can be used to reflect the stock market behavior. The correlation of stock prices can be examined by analyzing the stock market based on complex networks. This paper uses the stock data of Tehran Stock Exchange from March 21, 2014, to March 21, 2017, to construct its stock correlation network using the threshold method. With an emphasis on centrality in complex networks, this article addresses key economic and financial implications that can be derived from stock market centrality. Central industries and stocks are thus identified. The results of the analysis of stock centrality suggest that stocks with a higher market capitalization, a greater risk, a higher volume of transactions and a lower debt ratio (i.e. greater liquidity) are more central. These stocks attract more customers due to their attractive investment features and thus have a greater market influence. The review of the relationship between centrality and the growth of industries shows that an industry or a sector with greater economic growth has a higher centrality value and is positioned more centrally in the stock market network.

Introduction

One of the most important problems in modern finance is finding efficient methods for visualizing and summarizing stock market data. A significant volume of daily data is produced by the stock market, and this information is presented by thousands of plots and separately represents the price movement of each stock. When the number of stocks increases, these plots become more complicated to analyze [1]. Moreover, if there are multiple heterogeneous components, the stock market behavior becomes complex [2]. In addition, stock price fluctuations are not independent of each other and have a strong correlation with the business and industries to which the stock belong [3]. Based on recent investigations, the complex network method is highly recommended for visualizing and summarizing stock data and studying the correlation of stock prices [4,5]. A crystal clear image of the inner structure of the stock market can be presented by complex network analysis [2]. In this method, unlike classic cost–benefit methods, stock price variations are affected by group behaviors. Studying the structure of the stock market network helps explain the stock market behavior and the interaction among its factors. This method therefore challenges the independent variable assumption of the current linear analysis methods that are based on identifying the effects of several independent variables on a dependent variable [6]. The stock market network is constructed based on the correlation between stock price returns. Studying the correlation matrix has a long history in financial affairs and is the main foundation of Markowitz theory about securities market [7]. The correlation analysis of financial markets is an important issue for market policymakers and activists, such as portfolio managers, and is also crucial for risk management and asset allocation [8]. Various studies conducted to analyze complex networks in the stock market also confirm these findings. Brida et al. [9], Zhong et al. [10] and Zhao et al. [11] analyzed the topologic structure of the financial market network and proposed network analysis as a helpful guide for investors. Eberhard et al. [12] investigated the network properties of the stock market in Chile. Their results showed that the structure of the stock market network in Chile can affect the stocks’ transaction volume and return in the market. Sharma et al. [13] constructed the stock market network for India based on the correlation among market stocks using the threshold method and concluded that the network analysis of the Indian stock market can provide a better understanding of stock correlations in this market.