Electronic Nose Dataset for Classifying Rice Quality using Neural Network

Abstract

Rice is a staple food ingredient because it is the main food element for Indonesia and the world. However, the quality of rice can decline over time until it becomes expired or smelly and cannot be consumed. At present, the conventional method to distinguish between expired rice and not expired rice is still carried out by observing rice with the human sense of smell. However, this method is still considered ineffective because the human sense of smell can change due to changes in body health. In this case, this study uses an electronic nose (enose) and a machine learning neural network (NN) algorithm to detect rice consistency (expired and non-expired). The dataset was obtained from the e-nose by recording sensor information for 25 weeks by storing 48.486 total data and 2.017 data records for one week. The results of the classification using NN are with an accuracy score of 99.84%, the proposed method successfully classified rice quality.