Temperature: Higher temperatures increase AC usage, leading to more condensate production. Each degree above 25°C adds approximately 2L to the prediction.

Humidity: Higher humidity means more moisture in the air, resulting in more water recovery. Each percentage point above 50% adds 0.3L to the prediction.

Time Patterns: Historical data shows consumption peaks during afternoon hours and lower demand during night hours. The system learns these patterns over time.

Model Accuracy: The system continuously learns and improves predictions over time, currently achieving 93.4% accuracy.