Application Biplot and K-Medians Clustering to Group Export Destination Countries of Indonesia’s Product

Abstract

A good increasing export will yield foreign exchange to a country, and subsequently funding its country growth. In Indonesia, export is one of the biggest foreign contributors. As we can see that the countries Indonesia export to are more than 100, it is a must to group them based on their similarity. Biplot and cluster analysis are statistic methods which are used as tool to classify data based on variable explanatory. There are outliers in data acquired. Outliers are observation data which is appeared to be extremely different to the other data. Those data are identified by leverage method. in summary, this research applies K-Medians Clustering Method using Manhattan Distance to resolve outliers while grouping the countries based on their export data. The data contains export data of 182 countries in the year of 2017. R 3.5.1 software was used to calculate in this analysis. The clustering shows us that each continent form difference clusters. Asia has 4 clusters while the rest each has 3 clusters. In addition, we can conclude that several clusters have high value export of Indonesia for certain variables.