| dc.description.abstract | The assessment of food security as a measure of development is very important. Food security in a region has multidimentional characteristics that need to be analyzed. The purpose of this research is to implement the S-K algorithm (combination of Self organizing maps -SOM- and K-means algorithm) for data clustering and to gain data characteristics as the result of data clustering. The used data is the indicator for the food security from 30 districts in the provinces of Bali, Nusa Tenggara Barat, and Nusa Tenggara Timur. These data are the input for S-K algorithm. SOM clustering result is validated using Davies-Bouldin Index (DBI). Centroid and the number of cluster from SOM are utilized as the input for K-means algorithm, which is used to refine the final cluster. In this research, these data are also clustered by K-means algorithm with randomly generated initial centroids. The value of DBI results of SOM, S-K, K-means clustering has been compared and it is found that S-K algorithm has the minimum value of DBI. Thus, it is proved that the S-K algorithm gives good clustering results. Based on the data analysis, the districts in the Province of Nusa Tenggara Timur are categorized as the areas with food insecurity. Meanwhile, the districts in the Province of Nusa Tenggara Barat are included in the relatively food insecurity areas. Food security in all districts in the Province of Bali are satisfactory. | en |
