Grouping Madura Tourism Objects with Comparison of Clustering Methods

Madura has tourist attractions spread over four regencies, one of which is Bangkalan. Bangkalan has the most tourist attractions compared to other districts in Madura. There are twenty-one attractions in Bangkalan from three categories: religion, nature, and culture. And there are several indicators that influence visitor needs in determining tourist attractions, including Name of Tourism, Type of Tourism, Region, Gender, Age, Occupation, Education, and Marital Status, as well as three groups of high, medium, and low. Clustering is a data mining process that can be applied in various fields. The method used in this study is K-Means and Density-Based Spatial Clustering Algorithm with Noise (DBSCAN) methods. The K-Means method is a data mining method by grouping non-hierarchical data. At the same time, the DBSCAN method is a grouping method based on data density. This study compares the two methods to produce optimal solutions for classifying tourist objects. From several trials that have been carried out by comparing the two methods, the resulting accuracy value using a dataset of 21 attractions with ten criteria shows the accuracy value obtained by K-Means is superior to DBSCAN. The K-Means method has a higher Silhouette Index (SI) value of 0.6902 and k= 8.