OPTIMIZATION METHODOLOGY WITH PRINCIPAL COMPONENT ANALYSIS APPLIED IN PORTFOLIO: SIMULATED RESULTS IN GREEN COMPANIES

This research objectives to develop mathematical models for risk-return that consider simultaneously the linear and non-linear effects of the invested proportions and features of green companies based on their balance sheets/annual reports indicators. The mixture design of experiments is combined with Ward clus-tered PCA process variables for selecting the most promising companies through generated financial indicators (General Multivariate Indicators method) from the green companies. The green companies are the object of study for this empirical research, they were gathered from Newsweek magazine rank. The pro-posed Clustered Multilevel Optimization method has showed to be more robust and efficient than the all-other tested methods in this research. That means greater security and less risk to the investor.