Comparison Between Classification and Regression in Decision Trees
Aspect Classification Regression
Goal To predict a category or class label (discrete To
output).
predict a continuous value or numerical output.
Output Type Discrete (e.g., 'Yes/No', 'Class A/B/C'). Continuous (e.g., price, temperature, age).
Split Criteria - Measures like Gini Index, Entropy (Information
- Splits
Gain),
areordetermined
Chi-Squareusing
are used
metrics
to split
like Mean
the nodes
Squ
Leaf Node Value Represents the majority class of the data points
Represents
in that node.
the mean or predicted value of the data
Example Problems - Spam email detection.- Tumor classification- (malignant/benign).
Predicting house prices.- Forecasting sales reven
Evaluation Metrics Metrics like Accuracy, Precision, Recall, and Metrics
F1-Scorelike
are
MSE,
used.
RMSE, MAE, and R-squared are
Pruning Can prune based on class misclassification orCan
entropy
prunereduction.
based on reduction in error (e.g., MSE).
Visualization Final tree represents class labels at the leaf nodes.
Final tree represents continuous values at the leaf
Decision Boundaries Creates piecewise constant boundaries for class
Creates
separation.
piecewise constant boundaries for numeric
