The word "learning" in ML (Machine Learning) refers to the process through which computers analyze current data and learn new skills and knowledge from it. ML systems use algorithms to look for patterns in datasets that include unstructured and structured data, numerical and textual data, and even rich media like pictures, audio, and video. Because ML algorithms are computationally intensive, they need specialized infrastructure in order to operate at large sizes. The three fundamental kinds of ML are supervised ML, unsupervised ML, and reinforcement ML, which are discussed in this article. The supervised learning method is described, and it demonstrates how to utilize supervised ML by splitting data into training and testing, and how training all prior data aids in the discovery of the predictor. Unsupervised ML, which helps to divide categories into different clusters or groupings, is then addressed in this article utilizing techniques such as k-means and idea component analysis. Finally, this article looks into reinforcement ML, which uses the right behavior to maximize rewards.