소셜네트워크 기반 관광객 이동 예측을 위한 문법추론 히든마코프모델

Hidden Markov Model to Predict Tourists Visited Places Theo Demessance L´eonard de Vinci, Research Center, 92 916 Paris La D´efense, France College of Intelligence and Computing Tianjin University theo.demessance@edu.devinci.fr Chongke Bi College of Intelligence and Computing Tianjin University bichongke@tju.edu.cn Sonia Djebali L´eonard de Vinci, Research Center, 92 916 Paris La D´efense, France sonia.djebali@devinci.fr Guillaume Gu´erard L´eonard de Vinci, Research Center, 92 916 Paris La D´efense, France guillaume.guerard@devinci.fr Abstract—Nowadays, social networks are becoming a popular way of analyzing tourist behavior, thanks to the digital traces left by travelers during their stays on these networks. The massive amount of data generated; by the propensity of tourists to share comments and photos during their trip; makes it possible to model their journeys and analyze their behavior. Predicting the next movement of tourists plays a key role in tourism marketing to understand demand and improve decision support. In this paper, we propose a method to understand and to learn tourists’ movements based on social network data analysis to predict future movements. The method relies on a machine learning grammatical inference algorithm. A major contribution in this paper is to adapt the grammatical inference algorithm to the context of big data. Our method produces a hidden Markov model representing the movements of a group of tourists. The hidden Markov model is flexible and editable with new data. The capital city of France, Paris is selected to demonstrate the efficiency of the proposed methodology. Index Terms—Tourist behavior, Hidden Markov Model, Gram- matical Inference. I. INTRODUCTION In 2019, World Tourism Organisation UNWTO1 recorded 1.5 billion international tourists, 4% more than the previous year. Tourism is one of the most important areas of the world economy. It is also considered to be one of the fastest-growing industries in the world [1]. In this economic context, the understanding and knowledge of travel motivations and the anticipation of tourists’ behaviors play an essential role in tourism marketing. It can lead to the recognition of demand, to make targeted advertising [2], and help tourists to make decisions [3]. With the recent booming of digital tools and mobile in- ternet technology, alternative sources of data to understand tourism behavior have emerged. Users of social networks, like Tripadvisor, Booking, Facebook, Instagram, tend to 1International tourism Growth continues outpacing the global economy: edition 2020. share openly and frequently photos, reviews, recommendations and videos of tourist places. Thus, when users share photos or reviews, geographical information is included. These geo- located data represent tourism and sociological views [4], [5]. Analyzing the behavior of tourists represents an important challenge to better monitor their movement and spreading in a given area. We can subsequently adapt supply to demand, recommend a stay to a tourist, provide relevant information for the tourism industry and management. In this paper, we focus our reviews on the prediction of tourist movements. This paper addresses the problem of modeling and predict- ing the future movement of a tourist based on his present and past practices in a given area. Based on the geo-localized and temporal data information, we propose in this paper a model for predicting future tourist movement by analyzing the time sequences of places visited by a set of tourists. Our approach is to learn tourist’s practices from a set of temporal sequences of places, through various methods to handle the difficulties due to big data. The proposed method uses a machine learning method building a Hidden Markov model representing the whole data set. The model can produce predictions as a recommendation of future places to visit. The model can also be updated to adapt to new data. Our model is built from the whole data without reducing its size and extracting a mathematical model. Thereby, we propose a new algorithm for automatic learning of grammatical inference to reduce its complexity in the context of big data. Moreover, this algorithm is designed to maintain all behavioral possibilities on the data set. The principal contributions of our works are: 1) A method to establish sequences representing a unique stay of a tourist from a data set; 2) A new method of grammatical inference for processing very large data set; 3) A flexible and relevant decision-making tool to represent all tourists’ movements in a data set. This decision- arXiv:2511.19465v1 [cs.LG] 21 Nov 2025 making tool is able to predict future visits as recom- mended places to visit. To validate the method, the results will be compared to the data statistics of the data set. Our method is not compared to deep learning methods of the literature. In this paper, we will first describe in Section II the related work on network analysis. In section

This content is AI-processed based on open access ArXiv data.