A machine learning approach for coconut sugar quality assessment and prediction

• Main Authors: Alonzo, Lea Monica B., Chioson, Francheska B., Co, Homer S., Bugtai, Nilo T., Baldovino, Renann G.
• Format: text
• Published: Animo Repository 2019
• Subjects: Machine learning; Sugar—Quality control; Manufacturing
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/2774
• Institution: De La Salle University

This study presents a machine learning approach to accurately assess the quality of coconut sugar using RGB values. Python and scikit-learn were used to run the following machine learning algorithms: artificial neural network (ANN), stochastic gradient descent (SGD), k-nearest neighbors (k-NN) algorithm, support vector machine (SVM), decision tree (DT) and random forest (RF). Comparisons were made between the aforementioned machine learning algorithms by evaluating the accuracy and the average running time of each training model. Results of the study show that the SGD is superior in terms of accuracy but falls short to k-NN and SVC in terms of running time. In this fashion, a plot between the accuracy and the running time was made and it was observed that algorithms with higher accuracies correspondingly have also higher running times. By this very nature, experimental results show that the SGD holds merit in accurately assessing the coconut sugar quality, despite its expense in running time. © 2018 IEEE.