Electricity Consumption Prediction & Billing
Tarun Gangadhar Vadaparthi
Abstract
We model electricity consumption for a college campus using 12 years of meter readings (2012–2023) and estimate monthly bills. After cleaning and aggregating the dataset into quarter-wise records, we train linear and polynomial regression models. Forecasted units are priced via the applicable tariff slabs (energy charges, fixed charges, wheeling, and taxes). The approach achieves strong accuracy on held-out periods and demonstrates a practical path from forecasting to actionable billing estimates.
Results
