C4.5 Decision Tree Construction

C4.5 is an improved version of ID3. It addresses the limitations of ID3 and enhances its capabilities.

Features of C4.5:

1. Handles both discrete and continuous attributes.
2. Supports missing values (represented as ?; ignored in calculations).
3. Performs post-pruning to reduce overfitting.
4. Uses Gain Ratio instead of just Information Gain.
5. Robust for large and noisy datasets.

Algorithm : C4.5 Decision Tree Construction

Input: Training dataset T
Output: A Decision Tree

Steps:

1. Calculate Entropy_Info for the dataset.
2. For each attribute A:
   • Calculate Info_Gain(A)
   • Calculate Split_Info(A)
   • Calculate Gain_Ratio(A)
3. Choose the attribute with highest Gain Ratio.
4. Use it as the root node and split the dataset based on its values.
5. Recursively repeat for subsets with remaining attributes until:
   • All instances belong to a single class.
   • No more attributes to split.
   • Leaf node is reached.

Predict Job Offer Using C4.5