A Random Forest Approach for Predicting the Microwave Drying Process of Amaranth Seeds

Abstract

In this work, a model has been developed for the prediction of the fundamental variables of the microwave drying process of amaranth seeds, using the initial mass of seeds and the temperature of the process as input data. The model was developed by using the RandomForestRegressor classifier, which is found in the module sklearn.ensemble of the Python programming language. For the training and prediction of the model, the data of the measurements made of the drying time and energy consumption in the drying experiments carried out at three temperatures (35, 45, 55 ° C) in a domestic microwave oven were used, as well as the germination rate of the amaranth seeds obtained in the germination tests. The predictions made by the model have a precision of 99.6% for the drying time, 98.5% for energy consumption and 92.2% for the germination rate of the seeds.