This method of data analysis brings together supervised learning algorithms adapted to quantitative data. The objective is to learn (to find) the relation that binds a variable of interest of quantitative type to the other variables observed, possibly for the purpose of prediction. We use regression when the variable of interest is quantitative ("value" in a metric space). The metric is a notion of distance defined in space and often of "continuous value".
For example, one can try to predict the age of a user according to his behavior. Age is a continuous datum with the usual metric of real numbers (23 years old and 22 years old are 1 year apart). The simplest regression algorithms are of the linear regression type, while the most complicated of the least squares regression type, neural network, machine vector support, and so on.
Read our blog to learn more.
