A machine learning approach to tourism recommendations system

Chia, An (2025) A machine learning approach to tourism recommendations system. Final Year Project, UTAR.

PDF Download (5Mb)

Abstract

This project aims to develop a tourism attractions recommendation system by integrating machine learning recommendation algorithms. The main problem encountered when developing a powerful recommendation system is cold start problem, data sparsity and scalability problems. Cold start problem occurs when there is insufficient data about new users, new items or both. Data sparsity is a situation where there exists null value in the dataset, making it difficult to make predictions. Scalability problems arise when the system struggles to handle large volumes of data or a growing number of users and items. To overcome this problem, this project implements machine learning algorithms with collaborative filtering, content-based filtering and hybrid filtering approaches. Algorithms like Singular Value Decomposition, K-Nearest Neighbor and Co-clustering will be compared in this project. Model with the highest accuracy will be integrated into a tourism recommendation mobile application. A high portability and mobility mobile application will be developed by using React Native, and the dataset used in developing will be obtained from Google API. By developing this powerful recommendation system, travelers, tour guides and tourism agents will benefit by reducing their massive workload on planning trip itinerary.

Actions (login required)