Description of Invention
Around 30-40% of global energy consumption is attributed to the building sector, with over 80% occurring during the building's operational phase (e.g., heating, cooling, ventilation, lighting) [1]. Accurate forecasting of energy consumption is crucial for large buildings; however, many models overlook the significant impact of occupant behavior on energy use [2]. To address this issue, it is essential to collect data, investigate relationships, define patterns, and apply algorithms that process input data to produce reliable outputs based on these patterns [3, 4]. This project aims to bridge that gap by collecting synchronized environmental, occupancy, and power usage data from sensors, and using machine learning models—such as K-Nearest Neighbors (KNN), Support Vector Regressor (SVR), XGBoost Regressor, Random Forest Regressor (RF), and Light Gradient Boosting Machine Regressor (LightGBM)—to explore how occupant activities influence energy consumption through sensor inputs and occupant questionnaires.