Principal Component Analysis (page 1 of 5)

Principal Component Analysis (PCA) is a method used in pre-processing (dimension or feature reduction) to prep our data for analysis. We often use a PCA to reduce our data down to a manageable number of dimensions (features) such that we can train and test a machine learning model in a reasonable amount of time and in a cost effective manner. Reducing the feature set (i.e., dimension reduction) obviously comes at the expense of accuracy, but we "trade" a bit of accuracy for simplicity by removing extraneous features to make the models simpler to implement and interpret.