Neural Network Modeling for Forecasting Tourism Demand in Stopi\'{c}a Cave: A Serbian Cave Tourism Study

TL;DR

This study compares classical, machine learning, and hybrid neural network models for forecasting tourist visits to Stopića Cave in Serbia, highlighting the superior performance of the NeuralProphet model that incorporates seasonality, trends, and exogenous variables.

Contribution

Introduces a hybrid NeuralProphet model combining classical and machine learning techniques for improved tourism demand forecasting in environmentally sensitive sites.

Findings

NeuralProphet outperformed ARIMA and SVR in prediction accuracy.

Inclusion of seasonality and trend improves model performance.

Google Trends as an exogenous variable enhances demand modeling.

Abstract

For modeling the number of visits in Stopi\'{c}a cave (Serbia) we consider the classical Auto-regressive Integrated Moving Average (ARIMA) model, Machine Learning (ML) method Support Vector Regression (SVR), and hybrid NeuralPropeth method which combines classical and ML concepts. The most accurate predictions were obtained with NeuralPropeth which includes the seasonal component and growing trend of time-series. In addition, non-linearity is modeled by shallow Neural Network (NN), and Google Trend is incorporated as an exogenous variable. 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. The data provided insights into the tourist demand in Stopi\'{c}a cave and preliminary data for addressing the issues of carrying capacity within the most visited cave in Serbia.