Model Performance Comparison
AI-trained models ranked by accuracy for Medicine SKU: MED-001
Accuracy % - How often the model gets it right (higher is better). 95% = correct 95 out of 100 times.
MAE (Mean Absolute Error) - Average difference from actual sales in units (lower is better). MAE of 12 = typically off by 12 units.
MAPE % - Percentage error on average (lower is better). 8% = predictions are within 8% of actual sales.
💡 TIP: The Ensemble Model combines multiple models for the best overall accuracy!
📊 What This Chart Shows:
Prophet is leading with the highest accuracy score, meaning it makes the most accurate predictions for your medicine. The Ensemble model (shown in the "Best Choice" tab) combines the top performers for even better results.
🎯 Accuracy
How close predictions are to actual sales
⚡ Speed
How fast the model trains and makes predictions
🛡️ Reliability
How consistent the model is over time
✅ Best Choice: Ensemble Model
The Ensemble model combines the strengths of Prophet, LSTM, and XGBoost. It's 95.8% accurate and works well for most medicines. This is what we recommend for you!
If you need simple & fast:
Choose Prophet - It's the fastest and easiest to understand.
If patterns are very complex:
Choose LSTM - Best at learning difficult patterns.