Codification of Dendrograms Portfolio Based on Euclidean Distance Measure (A Comparison Between Different Methods of Hierarchical Clustering)

Abstract

Today analysis of financial markets as a part of the capital market and its impact on development and portfolio design and investment strategy of each country has become an important and most critical issue. The aim of this study was to investigate how the connection and distribution of stocks related to 30 large companies index of Tehran Stock Exchange and the effects of relationship between clusters of related stocks to every industry. In this study, using a variety of methods of hierarchical clustering, structure, classification and hierarchy of the stocks in the year 1393 reviewed. The results showed that With a focus on each of the hierarchical clustering methods and their implementation on the target stocks, were identified different clusters of stocks due to the similarity and economic relationships and also the key clusters and the vital stocks in the desired set were obtained. The results indicate that the choice best hierarchical clustering algorithm for clustering stocks depends on the desired purpose of cluster analysis and consideration of the advantages and disadvantages of each method.

Publication
Financial Knowledge of Securities Analysis , (10), pp. 89-105