Principal Component Analysis (PCA)

Reduce dimensionality and visualize your data.

Definition

PCA transforms a large number of correlated variables into a reduced set of uncorrelated principal components that capture most of the data variability. It is ideal for exploration and visualization of multivariate data.

When to use it

• Reduce dimensionality of a dataset with many variables
• Visualize groups or trends in multivariate data
• Eliminate multicollinearity before regression
• Explore the underlying structure of data

Requirements

• Continuous and correlated variables
• Standardized data (mean 0, SD 1)
• N > number of variables recommended

What StatsLab computes

• Eigenvalues and % variance explained
• Scree plot
• Loadings (variable contributions)
• Biplot (scores + loadings)
• Individual scores on components
• Correlation circle

Worked example

Context : PCA on 8 biological parameters measured in 120 patients.

Result : PC1 explains 38% · PC2 explains 22% · Total: 60%

Interpretation : Two components are sufficient to summarize 60% of the information. The biplot reveals that blood pressure and cholesterol are strongly correlated (same direction on PC1), while age and blood glucose define an independent dimension (PC2).

Run this analysis