Redefining Library Spaces: The Role of Smart Libraries in Enhancing User Experience

Abstract

Purpose: Research aims to transform traditional libraries into intelligent, data-driven smart libraries capable of adapting to user needs in real time. It addresses the limitations of existing feature selection methods in handling high-dimensional data and evaluating user experiences accurately and at scale. Methodology: A novel framework combining Enhanced Binary Particle Swarm Optimization (EBPS) with Adaptive Extreme Gradient Boosting (Adaptive-XGBoost) was proposed. Data were collected through sensor-based interactions, smart application usage logs, and structured user surveys. Preprocessing steps included normalization, missing value imputation, and one-hot encoding to ensure feature quality. EBPS was employed to optimize feature selection, improving the Adaptive-XGBoost model’s predictive performance. All phases of data processing, model construction, and evaluation were carried out using Python. Results: The EBPS-Adaptive-XG Boost model achieved superior user interaction rates compared to XGBoost and Adaptive-XGBoost individually with touchscreen interaction increasing to 89.3%, recommendation usage to 91.0%, study area utilization to 93.7%, and book check-in/out frequency to 96.4%. Navigation assistance requests also improved to 87.5%. Sensor-based evaluations showed notable accuracy gains across RFID (89.2%), BLE beacons (90.1%), touchscreen sensors (91.6%), PIR motion sensors (90.7%), and Temp, Humidity, Light Sensors (88.8%), with multi-sensor fusion achieving 94.4% accuracy. Conclusions: The proposed model effectively supports continuous evaluation and improvement of smart library services. It offers a scalable, intelligent solution for real-time user experience assessment and data-driven decision-making in next-generation library environments.