Machine learning is a type of AI that mimics learning and becomes more accurate over time without being explicitly programmed to do so. Think recommendation engines, a common machine learning application which gives you more tailored ads the longer it observes your clicking habits. The two most common approaches to machine learning training are supervised learning and unsupervised learning. Supervised machine learning requires a data scientist to train the algorithm with labeled inputs paired with desired outputs. For instance, a shape with three sides is labeled as a triangle.
The goal is that the algorithm can eventually take in new inputs and independently predict the correct label for the data. Supervised machine learning is ideal for binary classification, multi-class classification, regression modeling, and on-sembling. Unsupervised machine learning does not require a data scientist to provide labels and corresponding outputs.
Instead, the algorithm sifts through unlabeled input data and identifies patterns that it will use to group data, like how many sides a shape has. It's unsupervised because there are no teachers and no right or wrong outputs. The algorithm is responsible for analyzing the underlying structure in the data. Unsupervised learning is ideal for clustering, anomaly detection, association mining, and dimensionality reduction.
Supervised learning is the more common of the two machine learning techniques. It provides accurate results, but can be complex, time consuming, and expensive to compute. Unsupervised learning is less complex and takes place in real time, but is also less accurate.