Clusterboost: An Airbnb Recommendation Engine Using Metaclustering
Ari Nikhil Sai, Vuyyuri Bhavani Chandra, Shaik Shabeena, Marisetti Nandini, Ongole Gandhi · 2024
In a bustling city like New York, selecting an ideal accommodation can be overwhelming due to the multitude of available options on platforms like Airbnb. This paper introduces ClusterBoost, a recommendation engine designed to offer personalized accommodations by leveraging meta-clustering techniques. Using a comprehensive Airbnb dataset, multiple clustering algorithms, including K-Means, DBSCAN, and Agglomerative Clustering, are employed to group similar listings based on features such as location, amenities, and pricing. The system then applies an ensemble model to combine the strengths of these algorithms, ensuring a more accurate and refined recommendation process. The final step of Spectral Clustering further optimizes the recommendations by aggregating results from the individual clusters. Cosine similarity is used to match user preferences with available listings, providing tailored suggestions. By combining these clustering techniques, ClusterBoost enhances the recommendation process, offering personalized, user-centric accommodation suggestions that improve user satisfaction and streamline decision-making.