Exploring Youth Drug Use with Decision Trees

Tags: Classification | R | Random Forest

Link: GitHub

🔍 Overview

Modeled marijuana usage patterns among youth using decision trees and ensemble methods.

❓ Key Questions

• Which factors influence first-time marijuana usage?

• Can social influence predict future use?

🧪 Methodology

• Used bagging, boosting, and Random Forests

• Handled 3K+ survey records from NSDUH

📈 Results

• Accuracy: 89.3% (binary) and 89.6% (multi-class)

• Peer influence and friend opinion were strongest predictors