A Novel Machine Learning Approach for Multidimensional Dynamic Destination Recommender S earch System Employing Clustering using Optimization Techniques

Objectives: To find the most suitable destinations according to a selection of various dimensions by the user. Optimization techniques are applied to clustering with the help of various objective functions, to find optimal clusters for better recommendations. Methods: This approach uses a hybrid filtering system for recommendation with a weighted K-means clustering algorithm and uses optimization techniques to improve cluster formation recommendation accuracy. This study used a data set from Kaggle, which considers different city names, types, and significance. According to city names, using a geolocator object in Python gets latitude and longitude for precise clustering. Findings: The elbow method utilizes the K-means and weighted K-means clustering algorithm to determine the number of clusters. We use optimization techniques such as the Artificial Bee Colony (ABC) algorithm and Particle Swarm Optimisation (PSO) to improve the cluster formation recommendation accuracy, as demonstrated by the following results: The accuracy of K-means with ABC and PSO is respectively 85% and 88%, while weighted K-means with ABC and PSO is respectively 90% and 93%. Novelty: This study highlights how advanced optimization techniques like ABC and PSO enhance K-means and weighted K-means clustering accuracy, precision, and recall. The combination of weighted K-means and PSO further enhances performance, making it the perfect choice for tasks that demand high-quality clustering and recommendation systems. Compared to the K-means clustering algorithm with PSO optimization, the ratio of successfully recommended relevant destinations is 7% higher. Keywords: Recommender System, Clustering, Destination Recommender System (DRS), Machine Learning, Weighted K­means clustering Algorithm, ABC, PSO