Support Vector Regression Algorithm Modeling to Predict the Availability of Foodstuff in Indonesia to Face the Demographic Bonus

Abstract

The demographic bonus becomes a valuable phenomenon for Indonesian. One of the positive effects of this phenomenon is the increase of productive age proclaimed which will be the future of Indonesian economy. The agricultural sector plays an important role of the overall national economy which is indicated by an increase from year to year. However, the level of nutritional adequacy declined by a few percent each year due to an increase in the number of people who are not balanced by increased demand for food. In this case the government is expected to determine the policy priorities related to Demographic Bonus issues by predicting the future. Computing and data mining technologies play an important role in prediction cases by drawing conclusions based on regression lines. The technique is called Support Vector Regression, which is able to handle some cases of statistical data. Three determinant attributes used in this research are (1) Harvest Area; (2) Number of Harvest Production; and (3) Food Productivity, become the main reference for 714 data from 1998-2015 in 34 Provinces in Indonesia containing 7 types of crops. Three distribution data experiments conducted using K-Fold Cross Validation have the highest accuracy on Fold-1 with correlation coefficient value (R) of 92% with the smallest error value at fold-1 with MSE value of 14%. Predicted results show a decline in the number of food production in almost every province in Indonesia. From the experimental results, it is known that the biggest contributor of food products is Java Island, especially in East Java. Almost every kind of palawija plant, East Java plays an important role in the production of food needed in Indonesia.