Indoor Environmental Quality Prediction Using Hybrid Deep Learning and a Comprehensive Environment Index

Abstract

Maintaining a healthy indoor environment is crucial for a productive and well-balanced life. This study proposes a comprehensive indoor environment index (IEI) that integrates air quality, thermal, visual, and acoustical comfort indicators using sensor data. Major indoor pollutants (CO, PM2.5, and PM10), temperature, relative humidity, noise levels, and illuminance are combined through an analytic hierarchy process to formulate the IEI. A hybrid deep learning model based on a CNN-GRU architecture is then used to forecast indoor environmental states across four categories (severe, very poor, poor, and satisfactory). ANOVA and Tukey′s HSD analysis confirmed significant differences among these categories. The model was trained on 80% of the dataset and tested on the remaining 20%, with performance evaluated using precision, recall, F1-score, and AUC-ROC. The proposed approach achieved a mean F1-score of 0.96, demonstrating high predictive accuracy and reliability. These results confirm the robustness and reliability of the proposed model. The study demonstrates its potential for supporting accurate indoor environmental quality prediction and providing a foundation for informed building management decisions.