Improving Indonesian Coffee Quality Through Accurate Predictions of Coffee Beans Moisture Content Based on the Bioelectrical Properties and Artificial Neural Networks
Nazhif Ubaidillah (a), Hanna Fauziah Habibah (b) ,Wahyu Dwi Ristianingrum (b) , Intan Salsabila Putri (c) , Muhammad Hafizh Oktasa (b) Dimas Firmanda Al Riza (b*)

a)Environmental Engineering, Faculty of Agricultural Technology, Universitas Brawijaya, Jl. Veteran, Malang, Indonesia
b*)Agricultural Engineering and Biosystem, Faculty of Agricultural Technology, Universitas Brawijaya, Jl. Veteran, Malang, Indonesia
*dimasfirmanda[at]ub.ac.id
c) Biotechnology, Faculty of Agricultural Technology, Universitas Brawijaya, Jl. Veteran, Malang, Indonesia

Abstract

Indonesia is one of the major coffees producing country and currently known as the fourth largest coffee beans exporters. Along with the increasing of the coffee export quantity, the quality control is required to improve the marketable quality and competitiveness. There are several quality attributes required for green coffee beans to be able to be exported. Coffee quality commonly characterized by the proportion of the detective beans, moisture content, sensory properties, and also physical and chemical properties. The methods to evaluate these properties are varied. Among these properties, moisture content is important to ensure long storage time. The methods to measure content varies from destructive measurement to non-destructive measurement method. Spectroscopy based method, bioelectric, and other approach has been used to measure the moisture content. The prediction model that could be used also has a trend from a regular statistical model into machine learning or artificial intelligent methods. The advancement of hardware and software for instrumentation will enable easier and cheaper technology to provide measurement device widely for the farmer and other stakeholder. This method increases speed and accuracy in the measurement of coffee moisture content . This paper will give an insight into the current state of the green coffee bean quality evaluation method and the possibility to improve and implement appropriate technology that is needed by all stakeholders.

Keywords: Artificial Neural Network, Bioelectrical Properties, Coffee Beans, Moisture Content

Topic: Post-harvest Technology