Supervised vs Unsupervised Learning

In AI/ML domain, model development approaches can be segregated as "supervised" or "unsupervised". In supervised learning approach the data includes inputs and associated outputs. However the unsupervised approach doesn't require the input-output mapping. While such unsupervised models can be useful for discovering the underlying patterns, the supervised models help in developing predication models .

Regression & Classification models are generally supervised kind of models; and clustering models use unsupervised approach. Also many AI/ML techniques have related models in both the approaches. For example while Principal Component Analysis (PCA) is a unsupervised model used to identify the low dimensionality / uncorrelated components of a multivariate process. Corresponding Partial Least Square(PLS) approach is supervised modelling approach used to model input-output type of data.