Rice Shelf-Life Prediction Using Support Vector Regression Algorithm Based on Electronic Nose Dataset

Rice is an important commodity for Indonesian that is both relatively abundant and accessible in the country. However, after rice passing a certain shelf-life time limit, it is no longer fit for consumption nor sale. Conventionally, sample rice is taken from storage periodically to monitor its quality, relying heavily on human judgment in its process. To automate this, we proposed the method to predict rice shelf life to use a machine learning Support Vector Regression (SVR) based on the electronic nose (e-nose) dataset. Hence, the contribution of this study is using SVR to predict rice shelf-life based on the electronic nose signals. This study used the SVR model with two parameters, namely C and Gamma, and utilized the min-max data normalization method in its preprocessing stage. In terms of prediction accuracy, the results are evaluated using $R^{2}$ dan RMSE. Our test shows that the proposed method is considered accurate with an $R^{2}$ value of 0.9974 and an RMSE value of 0.3597.