Principal Component Analysis — Hands-On Tutorial

Dimensionality reduction through Principal Component Analysis (PCA).

Photo by carlos lugo on Unsplash

Principal Component Analysis or PCA is one of the most popular dimensionality reduction methodologies available to statisticians and machine learning practitioners. Before diving deeper into what this means, let’s talk about a few scenarios where we use such methodologies in our daily life, probably without even knowing.

  1. Search Engines: When we use Google or other websites to find the answer to a question, instead of matching our search queries word by word, they use methodologies to first reduce the complexity of our searches into smaller parts and then search for it — reducing the complexity results in faster results and is done through dimensionality reduction.
  2. Image Compression: Do you recall that time when you were trying to upload a picture to a website only to find out that the picture exceeded the maximum file size? After dealing with the resulting frustration, we then seek help from tools such as Photoshop to reduce the size of that image. What Photoshop or similar tools perform under the hood is called dimensionality reduction.
  3. Music: Music streaming services use various methodologies to reduce the size of the music that is being streamed to save us and…