Other models

In this section, we will see other models provided by Rattle in the Model tab, which aren't supervised learning. These methods are Linear and Logistic Regression, Neural Networks, and Survival Analysis.

Linear and Logistic Regression

Linear Regression is a statistical method to describe the relationship between one or more input variables and one or more output variables. The objective is to create a formula that models the relationships between input and output variables; in this way, we can use this formula to predict new observations.

Imagine you are the manager of a marina, your marina has a gas station and you need to predict the amount of gas oil you will sell during a summer day. On the Mediterranean coast, during the summer, the amount of gas oil sold is correlated to the temperature. The reason is obvious, on sunny days, the temperature rises and more tourists want to use their boats. The example is illustrated in this diagram:

Based on past experience, we know that on...