Automated Prediction of Preservation Index in Library Environments through Multiple Regression Analysis

This study presents a novel predictive model employing multiple linear regression to enhance the calculation of the Preservation Index (PI) for library collections. Building upon and expanding previous research on Internet of Things (IoT)-based temperature and humidity monitoring in library rooms, this investigation incorporates critical environmental variables: temperature, humidity, and light exposure. Data was collected over a three-month period using advanced IoT systems deployed in three distinct library environments, each presenting unique preservation challenges. The model's development involved a meticulous process of interpolating existing PI data to achieve finer gradations, followed by analysis through a sophisticated multiple regression framework. Results indicate that while light exposure's contribution to the model's predictive capability was minimal, the overall model precision was remarkably high, demonstrated by a low Mean Absolute Percentage Error (MAPE) of 2.615. This research underscores the predominant influence of temperature and humidity in PI forecasting while also providing nuanced insights into the subtle effects of light exposure on collection preservation. A correlation heatmap is included to elucidate the interrelationships among variables, offering a visual complement to the statistical findings and enhancing the interpretation of complex data interactions. This study contributes to the advancement of preservation strategies in library science, potentially informing more effective environmental control practices in diverse library settings.