In the world of data, most people who are not too familiar with data science would typically believe majority of it involves creating models and generating predictions or the response variable based on data. After finishing week nine of the immersive, this only one half of the story in the world of supervised learning. The other half is where there is no response variable involved known as unsupervised learning. Within unsupervised learning, many preprocessing features or X variables would be analyzed in terms of correlation and density between each other. The two methods involve K-Means Clustering and Principal Component Analysis. The concept of transfer learning applies PCA to linear regression models by examining coefficients.
There are so many applications of unsupervised learning even in the technical world. On Netflix, often you will see a list of movies/shows that are recommended for you based on the previous movies/shows you watched. Recommender systems for companies like Amazon, Hulu, Netflix are based off unsupervised learning with the purpose of providing the consumer with product and movie insights through complicated mathematical concepts in Linear Algebra including matrices and pivots. The movies are grouped by similarity or clustering. Assumingly, the same concept applies for the advertisements showing up on you web browser that link websites with potential products that you may be interested in. This is an example of how companies are starting to become data driven in today’s world.