MODELING OF SOYBEAN PRODUCTION IN INDONESIA USING ROBUST REGRESSION

Abstrak

Soybean is considered as a vital commodity for food security, but apparently the production is unable to compensate the rate of increase in community needs. Therefore, this commodity plays an important role in economic life and indirectly may affect the level of demand for other materials. So the availability of soybean plays a major role for economic stability. The aim of this paper is to construct a robust regression model for predicting the soybean production in Indonesia using M-estimation. Based on the data obtained from Susenas and BPS, we obtain a robust regression model for soybean production.
Harvested area and productivity have significant influence, while production of seed does not have significant influence on soybean production. The increment of one hectare harvested area and one quintal per hectare of productivity will increase 1.34 tons and 574 tons of soybean production respectively.
Keywords: Robust regression, M-estimation, Huber function, soybean.