In a mathematical frame of reference, machine learning algorithms take computational approaches and gather information from that instead of using a preset equation for its model. The long-term performance of the algorithm keeps getting better and better, as the volume of samples for learning accumulates. These machine learning algorithms seek out any type of regular patterns in the data, glean information from those patterns and then output predictions to generate improved decisions. There are two categories of ML methods: supervised learning and unsupervised learning. Supervised learning identifies patterns (and spits out predictive models) by incorporating input and output data. Supervised learning has two subcategories of classification and regression. Classification helps forecast distinct answers, such as whether a team will win or lose a game. Regression is applied when forecasting ongoing answers, such as trends in weather, etc. Unsupervised learning, in contrast, only determines patterns according to input information. This method can be beneficial when it's not known what one should be seeking. It can help with analysis of raw data.