Data_ScienceUse_Cases/Predictions/Random Forest

Folder Information

Diabetes Prediction - DONE

• No EDA
• Using RandomForestClassifier, SMOTE, sklearn, pandas libraries
• Result:
  • Attained Accuracy of 75.97%
  • Diabetes as Class 1 and No Diabetes as Class 0
  • Precision (Class 0) = 0.84, Precision (Class 1) = 0.64
  • Recall (Class 0) = 0.77, Recall (Class 1) = 0.75
  • F1-score (Class 0) = 0.80, F1-score (Class 1) = 0.69

Digital Marketing Expense Prediction - DONE

• Performed Exploratory Data Analysis (EDA)
• Using RandomForestRegressor, sklearn, seaborn, pandas, matplotlib libraries
• Dataset contains features:
  • Impressions, Reach, Video_plays, Link_clicks, Engagement, and Live_time
• Target variable: Amount_spent (in IDR)
• Result:
  • Attained R² of 0.86
  • live_time contributes 80% to the prediction of Amount_spent
  • Other features provide smaller but meaningful contributions
  • The model demonstrates robust handling of non-linear relationships and multicollinearity.