Property Recommendation Systems Based On Hybrid Filtering And Profile Matching

Housing is a fundamental need for every individual. However, the abundance of available property options often leads to information overload, complicating decision-making processes. To address this issue, this study develops a hybrid filtering-based property recommendation system that integrates content-based filtering (CBF) and knowledge-based recommendation systems (KBRS) with the Profile Matching method. This system aims to produce more personalized and relevant property recommendations while also mitigating the cold-start problem. During implementation, the system was evaluated using Black-Box Testing with Decision Table Testing, comprising a total of 300 test cases, 100 cases for each of the three user profiles: Single Individuals, Working Couples without Children, and Working Couples with Children. The testing results indicated that the system could classify the Single Individuals profile with an accuracy of approximately 61% and Working Couples with Children at about 75%, but only around 35% for Working Couples without Children. The confusion matrix analysis revealed frequent misclassifications of properties, indicating overlapping characteristics among segments. Overall, the system achieved only about 57% classification success. This suggests that although the hybrid approach can be applied by combining the advantages of CBF, KBRS, and profile matching, further refinements are needed, particularly in defining ideal profiles and weighting criteria, to improve the accuracy and relevance of recommendations in future work.