i have 1200 features highly correlated , and i want to reduce those number of features theso the best choice is use feature selection or dimensionality reduction? and which method is the best for my specific usein this case. feature selection can work with features highly correlated ??