Introduction to Machine Learning: Supervised vs. Unsupervised Learning

What is Machine Learning?

Machine Learning is the field that enables computers to learn without being explicitly programmed. The more data a model is trained on, the better it will perform. There are a couple different types of machine learning:

(1) Supervised machine learning: Is used most often in the real world. Supervised learning algorithms learn input to output mappings. You give your learning algorithms a number of examples to learn from as well as the right answer for each of those examples (right answer meaning the correct label y for a given input x). One important example of supervised learning is regression, which allows you to predict a number from a set of infinitely many possible outputs. Another example of supervised learning is classification, which predicts a category/class from a finite set of outputs.

(2) Unsupervised machine learning: Asks an algorithm to find patterns in the data all by itself. An example of this clustering (grouping related data points together). Another example is anomaly detection and dimensionality reduction (dataset compression).

(3) Recommender systems: Suggests/recommends additional products or content to consumers based on criteria such as past activity, purchases, race, gender, search history, etc.

(4) Reinforcement learning: Learns decision-making by learning the optimal behavior in a given environment to gain the maximum reward. Has huge applications in robotics.