Ranking Influential and Influenced Shares Based on the Transfer Entropy Network
Influence is a concept found in nature and society and is related to the interdependency among a set of objects. In the context of a stock market, the variation in price of shares can influence the variation in price of other shares, leading to influential and influenced shares. In this work we leverage the notion of transfer entropy to build a network of shares and pairwise directed influence that is used to rank the most influential and influenced shares. Classical network centrality metrics such as PageRank and HITS are leveraged to rank the nodes. We apply our methodology to the shares in the greater stock market in Brazil, we rank nodes to find source and destination of influence in that market, while also comparing the different rankings and their correlation with traded volume.