Implementing AI-driven Recommendation Systems to Enhance Digital Library User Engagement

Thesis Abstract

Digital libraries have become pivotal repositories of knowledge, yet user engagement remains a persistent challenge due to limited personalised recommendations tailored to individual user preferences. This study addresses the critical need to harness artificial intelligence (AI) to develop and implement recommendation systems that can significantly enhance user interaction and satisfaction within digital library environments. The primary aim is to design, deploy, and evaluate an AI-driven recommendation system tailored to digital library users to improve their engagement levels. Specific objectives include identifying key factors influencing user engagement, developing an intelligent recommendation algorithm based on collaborative filtering and content-based filtering techniques, and assessing the system’s effectiveness through user feedback and behavioural analytics. A mixed-methods research design was employed to achieve these objectives. Quantitative data were gathered via a structured survey administered to a stratified random sample of 400 active digital library users across public and academic libraries. The survey instrument measured variables such as user satisfaction, perceived usefulness, and engagement frequency, guided by the Technology Acceptance Model (TAM) and the Unified Theory of Acceptance and Use of Technology (UTAUT). Additionally, system log data detailing user interactions with the recommendation engine were collected to facilitate behavioural analysis. Qualitative insights were obtained through focus group discussions involving 30 participants, which aimed to explore user experiences, expectations, and perceived barriers associated with the recommendation system. Data analysis involved descriptive statistics to profile user demographics and engagement patterns, followed by inferential techniques such as multiple regression analysis to determine predictors of user engagement. The effectiveness of the recommendation system was evaluated through pre- and post-implementation comparisons using paired t-tests, supported by feature importance analysis via Random Forest algorithms. Thematic analysis of focus group transcripts provided contextual understanding of user perceptions, preferences, and concerns regarding AI-driven recommendations. Expected key findings include a statistically significant increase in user engagement metrics post-implementation of the recommendation system, with notable improvements in session duration, return rate, and overall satisfaction. Regression analysis is anticipated to reveal that perceived usefulness and ease of use are the strongest predictors of engagement, aligning with TAM and UTAUT frameworks. Focus group insights are predicted to highlight both the benefits of tailored recommendations and concerns related to algorithmic bias, data privacy, and system transparency. The study contributes to the existing body of knowledge by empirically demonstrating the impact of AI-driven recommendation systems on digital library user engagement, providing a comprehensive model that integrates technological and user-centered factors. It advances understanding of how intelligent algorithms can be effectively embedded within digital library platforms to foster sustained user interaction. Methodologically, the research offers a replicable approach combining quantitative behavioural analysis with qualitative user perceptions, adaptable to various digital library contexts. The main conclusion emphasizes that implementing AI-driven recommendation systems significantly enhances user engagement, with improvements contingent upon addressing user trust and system transparency. The study recommends that digital libraries integrate AI-based recommendation engines with user privacy safeguards, incorporate user feedback mechanisms for continuous system refinement, and provide proactive user education on algorithmic functionalities. Future research should investigate long-term effects of personalised recommendations on user loyalty and delve into AI ethics within digital library environments, including algorithmic fairness and bias mitigation strategies. Overall, this study underscores the transformative potential of artificial intelligence to redefine user experiences and foster more interactive, personalized digital library services globally.