Select test properties using inform

Select test properties using information gain metrics on each node of the tree, a measure called the goodness metric for attribute selection measures or splits. Select the property with the highest information gain as the test property for the current node. The property is the property with the highest degree of differentiation in a given collection. This information theory approach minimizes the number of expected tests required to classify an object and ensures that a simple tree is found. Set S is a collection of s data samples, assuming that the class label property has a m different class Ci (i,""m). The set si is the number of samples in class Ci. The desired information required for the classification of a given sample is given in the following formula: