ID3 and C4.5 algorithms [40 mn of reading - published 6/6/2006 7:48:47 PM - Target : Confirmé]

Author

Thibaut DELATTE Student-Engineer Supinfo Canada SUPINFO graduate year  2007    Write to the person    All projects of the same author    Mini-CV of the author

Resume

The purpose of the following article is to present two of the algorithms most used in the field of constructions of graphs of induction, and more generally in the field of data mining : they are the algorithms ID3 and C4.5.

Sommaire

• Introduction
• 1 General presentation
• 2 Construction of the tree
• 2.1 Rules
• 2.2 Principle of construction of the graph 
• 2.3 Reminder on the notations
• 3 Discriminating capacity   
• 3.1 The criterion for ID3
• 3.2 Entropy of a top   
• 3.3 Uncertainty on a partition   
• 3.4 Profit of uncertainty   
• 3.5 The criterion for C4.5    
• 4 Further
• 4.1 Conditions of stops for ID3 and C4.5
• 4.2 Pruning in C4.5   
• 4.3 Continuous attributes   
• Conclusion