Core Concepts
Modeling tourist demand for Stopića Cave in Serbia using classical ARIMA, machine learning SVR, and hybrid NeuralProphet methods to gain insights for sustainable management and conservation.
Abstract
The study aims to model the tourist demand for Stopića Cave in West Serbia, which has experienced a significant increase in visitor numbers in recent years. The analysis includes a comprehensive time series dataset of monthly visitation figures from 2010 to 2023, allowing for the exploration of visitor trends and the application of various forecasting methodologies.
The key highlights and insights from the study are:
The time series exhibits a strong seasonal component, with yearly peaks during the summer months (July and August) when visitation is highest, and lower numbers during the winter. There is also an overall growing trend, especially in the second half of the study period.
An interesting event was the highest number of visits recorded in the summer of 2020, during the COVID-19 pandemic, when domestic tourism in Serbia increased due to travel restrictions.
The Google Trend index for the keyword "Stopića pećina" (Serbian for "Stopića Cave") shows a strong correlation with the actual number of visitors, indicating that online search behavior is closely linked to visitation patterns.
Three forecasting models were evaluated: classical ARIMA, machine learning Support Vector Regression (SVR), and the hybrid NeuralProphet method. The NeuralProphet model, which incorporates the seasonal component, growing trend, non-linearity modeled by a shallow neural network, and Google Trend as an exogenous variable, provided the most accurate predictions, with a Root Mean Square Error (RMSE) of 1726.88, outperforming the other models.
The NeuralProphet model also provides interpretable outputs, including the significance of lags for both the auto-regressive and exogenous variable components, which can help policymakers and managers better understand the model and make informed decisions.
The insights into tourism demand patterns, seasonal variations, and visitation trends are crucial for assessing the carrying capacity of Stopića Cave and developing effective management strategies to ensure the long-term sustainability of this environmentally vulnerable destination.
Stats
The number of visitors to Stopića Cave increased from around 6,000 in 2010 to over 16,000 in 2023.
The Google Trend index for the keyword "Stopića pećina" reached a peak value of 100 in the summer of 2020.
Quotes
"Modeling tourist demand represents great importance for management structures and decision-makers due to its applicability in establishing sustainable tourism utilization strategies in environmentally vulnerable destinations such as caves."
"Predicting the increase in tourist demand can be of great importance for management structures, because it can indicate the need to implement certain measures to prevent overexploitation and over-tourism."