SOCR ≫

DSPA ≫

Topics ≫

Data Science and Predictive Analytics (UMich HS650)

Assignment 8: Decision Tree Divide and Conquer Classification

1 Explain these concepts:

• Information Gain Measure
• Impurity
• Entropy
• Gini

2 Decision Tree Partitioning

Load the SOCR Neonatal Infant Pain score data and follow these steps:

• Collect and preprocessing the data, e.g., data conversion and variable selection.
• Randomly split the data into training and testing sets.
• Train decision tree models on the data using C5.0 and rpart.
• Evaluate and compare the two models.
• Tune the rpart parameter and repeat the evaluation and comparison again.
• Assess the prediction accuracy and report the confusion matrix.
• Comment on different aspects of the prediction performance.
• Use various impurity measures and re-estimate the models.
• Try to use the RWeka package to train decision models and compare the results.
• Try to apply Random Forest and obtain variables importance plot.

SOCR Resource Visitor number