Tourism Recommendation system

Abstract

Tourism today has evolved into a widely accessible activity, with large numbers of people participating in travel and leisure trips for enjoyment and relaxation. However, travelers often face difficulties when deciding which destination to visit, where to stay, the associated travel costs, and the points of interest that best match their personal preferences. Considering these challenges, this project proposes the development of a webbased recommendation system that suggests suitable tourist attractions to users. For building the dataset, information was collected through web scraping from holidify.com, resulting in a dataset containing approximately 100,000 attributes related to various tourist destinations. The recommendation system generates suggestions based on two main factors: similarities among users and similarities in content. Collaborative Filtering is used to recommend destinations by analyzing the preferences of users with similar interests, while Content-Based Filtering suggests locations by examining similarities in the characteristics of different places. The machine learning model was trained using the collected dataset and later serialized (pickled) so that it could be efficiently integrated with the front-end interface of the website. Additionally, all relevant data and processed information were stored and managed using a MySQL database to ensure efficient retrieval and system performance.