Principal Component Analysis Explained

Datasets often contain far more columns, or features, than are actually necessary to understand the underlying pattern.

Principal Component Analysis, commonly called PCA, reduces this complexity by combining correlated features into a smaller set of new variables, called principal components.

These components are constructed so the first captures the largest possible variation in the data, the second captures the next largest, and so on.

Think of a dataset with fifty features describing a product, many of which overlap in meaning, like length, width, and overall size. PCA can compress this information into just a few components, without losing much meaningful detail.