Forecasting Tourism Revenues in Tanzania: A Comparative Analysis of Support Vector Machine and Artificial Neural Network Performance
Abstract
The tourism industry is a vital pillar of Tanzania's economy, necessitating reliable revenue forecasts for effective policy and investment decisions. This study analyses and compares the performance of two machine learning models, Support Vector Machine (SVM) and Artificial Neural Network (ANN) in forecasting Tanzania's tourism revenues. Based on historical annual data from 1970 to 2024, both models were developed, and their performance was evaluated using the Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) metrics. The comparative analysis shows that the Support Vector Machine model outperformed the Artificial Neural Network model, as evidenced by lower error rates. The findings suggest that SVM is a highly effective predictive tool for this domain, offering significant practical implications for the strategic management of Tanzania's tourism sector.
