HYBRID RECOMMENDER SYSTEM FOR INDOOR AND ORNAMENTAL PLANTS: AN AHP DRIVEN APPROACH

Abstract

Urban gardening in modern cities faces challenges such as limited space, pollution, seasonal variations, and lack of user knowledge about suitable plant species. Sustainable ornamental plant selection is vital for enhancing green spaces, improving environmental quality, and promoting eco-friendly lifestyles. However, traditional approaches often fail to integrate both environmental and user-preference-based factors, making plant survival and sustainability difficult in urban localities.The purpose of this study is to design and evaluate an intelligent recommender system for selecting indoor and ornamental plants tailored to urban environments. The study aims to balance ecological sustainability with user convenience by analysing plant suitability through multi-criteria decision-making. Specific objectives include identifying the most consistent and reliable recommendation approach while addressing limitations such as cold-start problems and inconsistent rankings in existing systems. The research employs the Analytic Hierarchy Process (AHP) combined with three recommendation techniques: Content-Based Filtering, Collaborative Filtering, and Hybrid Recommendation. Decision criteria such as sunlight requirement, seasonal suitability, number of plants, and growth rate were applied to evaluate four alternatives—Money Plant, Snake Plant, Air Plant, and Peace Lily. Pairwise comparisons, decision matrices, and consistency ratio analysis were used to determine the robustness and reliability of each method.Findings reveal that the Hybrid Recommendation method provided the most balanced and consistent results, effectively integrating both user preferences and plant attributes. Content-Based Filtering was strong in specific environmental criteria, while Collaborative Filtering proved valuable in user-centric contexts but struggled with sparse data. The study concludes that a hybrid approach is the most robust for urban plant recommendation, enabling personalized and context-aware solutions. The implications extend to sustainable urban landscaping, improved plant survival rates, and the development of smart gardening technologies. Recommendations for future research include refining hybrid models to reduce inconsistency and integrating real-time environmental monitoring for dynamic plant selection. This research contributes significantly to sustainable urban greening and practical household gardening.