A simple and efficient Inductive Learning algorithm
SILEA (a System for Inductive LEArning) is an efficient inductive learning algorithm for rule extraction. SILEA is a covering algorithm which extracts IF-THEN rules from a collection of examples in a reliable way.
SILEA employs a feature selection technique similar to that of Sequential Forward Selection to decrease the number of attribute-value
pairs that are to be considered. Sequential Forward Selection prioritizes certain attributes over the others by an objective function. Similarly, SILEA makes sure that these biases are made on attributes with higher metric values than the other attributes. The metric used for this purpose is the entropy measure. This approach allows SILEA to avoid consideration of enormous amount of possible combinations for each iteration.
The extraction process that SILEA employs eliminates unnecessary comparisons needed for rule extraction. It extracts all possible rules for each considered combination and selects the most classifying ones among them. It excludes the rules which might become obsolete due to the existence of more classifying rules. This approach assures the extraction of the most general rules for the considered combination of attributes.