Two main tasks of machine learning are regression with specific functions and classification of data into separate classes. Regression is a mathematical method that fits data with a curve, i.e., it passes an optimal curve through a given set of data. In linear regression we use a linear function to provide the best fit to the data. This line does not pass exactly through most of the points as it tries to make an interpolation through them. Nonlinear functions can also be used in regression, but the specific choice depends on the data characteristics and effort to avoid overfitting. Once the fitting function is determined, it can be used in the prediction of new data points. Classification works more with classes that are special groupings of data depending on specific features. Its aim is to learn how to separate the available data into separate classes, and then the model is presented with new data to be able to tell to which class they belong.