Analysis of Personalized Tourism Recommender Systems

EasyChair Preprint 9628

Abstract

Traveling is one of the most important leisure pursuits in life. While on vacation from one's regular routine, one experiences a new culture, new food, and new sights. As a result, recommending tourist destinations based on the user's interests may enable an individual to have incredible new experiences. Currently, different types of recommendation systems are used in the industry. Our primary goal is to compare all of the methods and determine which one is the best recommendation system for tourism. In this paper, we investigated various recommendation system methods used in tourist recommendation systems. The goal of this research is to identify all of the recommendation systems used in the industry. We tested which method would be the best for tourism recommendation by implementing them on a common dataset. Since content-based is item-based and collaborative filtering is user-based, these two alone are insufficient to produce an accurate result; therefore, a deep learning model is required in addition to these two. According to the parameters and factors considered for tourism, we discovered that the hybrid is the best option. Additionally, no recommendation system considered personality as a factor, so we are currently developing a recommendation system that does

Keyphrases: TF-IDF Equation, collaborative filtering, content filtering, cosine similarity, hybrid method

@booklet{EasyChair:9628,
  author    = {Dipit Sharma and Eklavya Banwala and Imane Elouaghzani and Rahul Katarya},
  title     = {Analysis of Personalized Tourism Recommender Systems},
  howpublished = {EasyChair Preprint 9628},
  year      = {EasyChair, 2023}}