What is dimensionality reduction in machine learning?

Updated May 16, 2026

Short answer

Dimensionality reduction is the process of reducing the number of input features while preserving essential information.

Deep explanation

It transforms high-dimensional data into a lower-dimensional space either by selecting a subset of features or by creating new composite features. Techniques like PCA or t-SNE aim to retain variance or neighborhood structure while removing redundancy and noise.

Real-world example

Used in image compression where thousands of pixels are reduced to a smaller feature set for faster processing.

Common mistakes

  • Confusing dimensionality reduction with feature selection only.

Follow-up questions

  • Is dimensionality reduction always lossless?
  • What are the two main types of dimensionality reduction?

More Dimensionality Reduction interview questions

View all →