<![CDATA[Tourism has great potential in driving economic growth through job creation, income enhancement, and positive impacts on various related sectors such as handicrafts, accommodation, and transportation. Garut Regency, located in the southern part of West Java Province, is increasingly recognized for its natural tourist destinations that remain unspoiled and attract many visitors. However, the surge in visitor numbers at these destinations has posed several challenges, including overcrowding that reduces comfort and safety, as well as a decline in service quality due to high demand. Inadequate infrastructure, such as transportation and parking facilities, is also an issue that needs to be addressed. To assist the local government in preparing for future increases in visitor numbers, this study utilizes the Linear Regression algorithm to predict the number of tourist visits to Garut Regency. This algorithm is chosen for its ability to measure the relationship between the dependent variable (number of visitors) and independent variables (factors influencing visits). Data collection is carried out by grouping the number of visitors based on tourist categories, resulting in more accurate and relevant prediction models. The research findings show that the linear regression model can generate predictions with a Mean Absolute Error (MAE) of 11,406.37, Mean Absolute Percentage Error (MAPE) of 6.449, Mean Squared Error (MSE) of 282,815,506.30, and Root Mean Squared Error (RMSE) of 16,817.12. The R-squared (R²) value of 0.9346 indicates that the model can explain approximately 93.46% of the data variance, demonstrating good predictive performance. However, the relatively high MAPE value indicates inconsistencies in the dataset, likely caused by very small or zero actual values. This prediction is expected to assist the Garut Regency Tourism Office in strategic planning and decision-making, such as infrastructure preparation, service quality improvement, and tourism promotion planning. This study also opens up opportunities for further development using other prediction algorithms to achieve more optimal results.]]>
