You are building linear prediction models (like linear or logistic regression) for a dataset with many features. Exploratory analysis shows many features are highly correlated, which can destabilize the model. What should you do to reduce the negative effects of having many correlated features?

Principal Component Analysis (PCA) is an effective technique for dimensionality reduction that transforms correlated features into a new set of uncorrelated features called principal components. This process mitigates multicollinearity, which can destabilize linear models and make coefficient interpretation difficult. By reducing the number of features while retaining most of the variance, PCA improves model stability and often generalization performance. Applying one-hot encoding is for categorical features and doesn't address correlation. Combining features with matrix multiplication is not a standard or effective method for handling multicollinearity. Computing Pearson correlation coefficients is a diagnostic step, not a solution to the problem of correlated features.