Applying the agglomerative method in hierarchical clustering for the medium-sized companies listed on the Warsaw Stock Exchange

Abstract

The purpose of this article is to use a hierarchical algorithm to reduce the number of companies in stock exchange portfolios, together with the identification of the most and least profitable groups of the companies. To prepare the research, the author decided to use a hierarchical clustering method to segment mWIG40 index entities. The conducted research contributed to the knowledge of the segments appearing on mWIG40 index and the profitability of the obtained clusters in the analyzed period. It was concluded that the hierarchical clustering method can divide the entities from mWIG40 index into six segments. The obtained groups differed from each other in terms of the analyzed features. Moreover, it was found that it was possible to identify more and less profitable segments in terms of the rate of return. What is more, only one segment was characterized by a higher rate of return than the benchmark. The findings can help investors to make better decisions during their investing process. In addition, the results can help companies to map their business in the market.