Reza Hosseingholizadeh; Mahmood Alborzi; Abbas Toloie Eshlaghy; Hamid Zargham Boroujeni
Abstract
The beliefs of the stakeholders who provide tourism services in destinations play an important role in the sustainability of destinations. The values of variables affecting these beliefs are always ambiguous, and the causal relationships between them are uncertain. The reality of opinion dynamics ...
Read More
The beliefs of the stakeholders who provide tourism services in destinations play an important role in the sustainability of destinations. The values of variables affecting these beliefs are always ambiguous, and the causal relationships between them are uncertain. The reality of opinion dynamics reflects even more complexity. Therefore, the present study used fuzzy logic, Bayesian belief networks, and opinion dynamics models to investigate a complex topic. The models were applied to the data obtained from the Iranian stakeholders in a bid to attract Chinese tourists. The data was collected by administering the questionnaire to 540 stakeholders during 2019, 2020 and 2021. Finally, the research provided not only an agent-based simulation to visualize trends and their evolution or co-evolution, but also the possibility of predicting defective and desirable trends with the ultimate aim of preventing or reinforcing them.IntroductionDespite the apparent regularity of the tourism process, ambiguity, uncertainty, and dynamism in the host community complicate the control of the process. It is ambiguous because different tastes are applied to it (Liu et al., 2021). It is uncertain because the stakeholders have different beliefs (Merigó et al., 2012). It is dynamic (Sainaghi & Baggio, 2017) because it is affected by environmental changes such as diseases (Soltan-Alinejad & Soltani, 2021), natural disasters (Rosselló et al., 2020), information technology (Xiang, 2018), and environmental issues (Saviolidis et al., 2021). Accordingly, conventional statistical tools are ineffective in capturing the whole complexity at work (Sun & Müller, 2013). In this respect, the current study used fuzzy logic, conditional probability, and diffusion processes on complex networks in order to model ambiguity, uncertainty, and dynamism, respectively. The fuzzy membership functions were used for the initial processing of the collected data (Serrano-Guerrero, 2021). The study utilized Bayesian belief networks to measure the community’s parameters in the current situation (Rohmer, 2020). Moreover, opinion dynamics was employed to predict future trends (Ureña et al., 2019). Specifically, the structural and parametric training of Bayesian belief networks was used to model the current beliefs of stakeholders in tourism destinations (Rohmer, 2020). The developed Bayesian belief network is able to make predictions, prescriptions, and causal inferences about the stakeholder-related variables of the community (Nadkarni & Shenoy, 2001). Therefore, various parameters of the community can be estimated by Bayesian belief networks (Gallardo, 2022). Stakeholders’ opinions are dynamic because they are influenced by contextual factors and other stakeholders (Steils et al., 2021). The study also intended to simulate the dynamics. There are many variables in stakeholders’ interactions that are influenced over time. Each stakeholder, also called the agent, has an opinion on each variable, so in interactions, stakeholders are influenced by their own opinions and those of others (Cape et al., 2018).Materials and MethodsThe current study aimed to create a structure for the beliefs of tourism stakeholders on the basis of tourists’ wishes and needs. Therefore, Bayesian belief networks, as a kind of causal network based on conditional probability theory, were used in this research. In the next step, the research employed opinion dynamics which is a subset of agent-based methods in the field of simulation. Agent-based methods, though recommended, are less addressed in the field of tourism. The data was collected by administering the questionnaire to 540 stakeholders in Iran during 2019, 2020, and 2021. The population of the stakeholders consisted of five classes including 265 hotels and accommodation centers, 170 tourism companies, 85 bus companies, 15 airlines, and 5 railway companies. The stratified sampling method was used to select the research participants.Results and DiscussionBy generating a joint probability distribution function for the variables governing the beliefs of the Iranian tourism stakeholders, the research integrated the beliefs into a coherent mathematical structure. In practice, this equation represented a kind of causal network structure that can answer different questions with the advantage that it can be retrained to be used to create a more updated distribution function when new data is entered. Relying on the agent-based method of opinion dynamics, the study simulated the future behavior of tourism stakeholders. The simulation showed what behaviors can be useful in attracting tourists, what behaviors are not in line with protecting the environment, and what behaviors can be suitable or unsuitable for attracting female tourists.ConclusionThe research was designed to use the methods separately in order to avoid increasing the complexity of the research process due to the simultaneous use of diverse and powerful tools. Therefore, it is possible for researchers to use each of the methods independently. In addition, the obtained results are useful, diverse, and practical, and suggest the suitability and further use of agent-based methods (e.g., opinion dynamics) in the tourism research
