In Partial Least Squares (PLS) regression, what type of combinations are sought?

In Partial Least Squares (PLS) regression, the primary objective is to find linear combinations of the input variables (also known as independent variables) that can effectively capture the variance in the response variable (dependent variable). This approach is particularly beneficial when there is high multicollinearity among predictors or when the number of predictors exceeds the number of observations, which can complicate traditional regression methods.

The linear combinations of inputs are generated in such a way that they maximize the covariance between the inputs and the outputs. PLS effectively reduces the dimensionality of the data while retaining as much information as possible, making it suitable for scenarios with complex relationships or large datasets.

In contrast, the other options focus on different aspects of regression or data processing that do not align with the underlying principles of PLS.