A classification based on support vector machines for monitoring avocado fruit quality

Abstract

Scientifically, the efficiency of a method refers to its power to best predict/calculate based on an evaluation following a certain process within the current scenario, parameter and/or data. For a good prediction, the most appropriate approach(es) to a problem should be considered and the related tests should be done reliably. Practical studies in the field of food safety and fruit quality are critical, with the accuracy, speed and economic parameters of the methods used being of particular importance. In this study, for the first time in literature an Arduino-based temperature and gas monitoring system (called e-nose) is used to monitor the decay of avocado fruit in a controlled experimental environment and support vector machines, a machine learning method, are used to detect (classification) the decay. In this study, test and validation success of over 99% was achieved with very few training-data for classification. The obtained results are encouraging in terms of the detection results of the developed e-nose and the method used to determine the level of decay in other fruit in cold storage.