MVLSC Home · Issue Contents · Forthcoming Papers
Unsupervised Classification Model Using Environmental and Financial Performance Attributes: Case Applied in Latin American Countries
Rodríguez García, Martha De Pilar and Lara Velázquez, Pedro
This research aims to determine the optimal clustering based on the classification of environmental and financial performance variables. The study was conducted using a dataset spanning from 2010 to 2017, consisting of 102 listed companies from the Mexican, Brazilian, Colombian, and Chilean capital markets. As a result, the number of companies included in each classification varied. Using profitability-based metrics, our findings indicate that there are no discernible advantages to being classified as NC company. Additionally, the mean and return values of both groups are highly similar. By employing risk metrics, the algorithm classifies the data according to countries, leading to a natural classification. Conversely, when considering eco-efficiency metrics, we observed the formation of two distinct groups. Therefore, it can be concluded that the algorithm effectively identifies a specific category of controversial companies by incorporating environmental performance variables.
Keywords: Financial performance, environmental performance, risk performance, unsupervised classification models, clustering and latin american countries