Improving Partial Least Squares regression based on meta-heuristic optimization algorithms

A popular method that combines and generalizes features from multiple regression and principal component analysis is called partial least squares regression (PLS). When we need to predict a collection of dependent variables from a very big number of independent factors, PLS is especially helpful. However, the performance of PLS depends on the number of components. Meta-heuristic algorithm implementation has grown in popularity among researchers. In this study, a Meta-heuristic algorithm is proposed to determine the number of components. The suggested approach will effectively assist in identifying the appropriate number of components with a strong forecast. The proposed approach is contrasted with two well-known approaches. The experimental results provide a thorough demonstration of the suggested method's superiority in prediction ability.