An Analysis and Design of Downstreaming Decision System on Palm Oil Agroindustry Based on Multilabel Classification

Abstract

The decision making improvement to enhance downstreaming in palm oil agroindustry is needed due to limited variety of infrastructures and their available market. Based on the fact, it requires correct labelling with multi attributes of an integrated sustainable palm oil downstreaming process of palm oil agroindustry in effort to support appropriate decision in downstreaming palm oil in the whole. In this paper, the result of business process model notation (BMPN) shows that the downstreaming decision form quantitave multilabel classification. The multilabel dimensions consists of technology availability, export taxes, competitive advantage, and product cost attributes. This research main contribution is to produce an integrated decision making rule using the multilabel decision tree or known as Bayesian Chain Classifiers (BCC) method that can classify the palm oil downstream decisions represented as a descriptive model. The result of downstreaming decision making model is obtained by considering four downstream aspects that are classified into multilabel objective and generate selected 11 appropriate decision rules from 139 datasets to provide available alternatives comprehensively for decision making. The rules performance shows an adequate accuracy but good precision, concluding that the approach suit to improve the downstreaming decision.