Development of Predictive Models using Machine Learning Algorithms for Food Adulterants Bacteria Detection

Abstract

One of the necessities of human to survive is food and meat is one of mainly consumed food by humans. Thus, a level of quality of food is a must to be safely consumed. There have been some cases of adulteration of meats, which can cause harm to consumers. Adulteration can lead to bacteria contamination which are difficult to determine the presence of bacteria without an instrument or food laboratory tests. Nowadays, the idea of applying machine learning in the field of food microbiology is becoming a trend. And one of these applications is on detection and classification of bacteria in food products. Hence, this study aims to apply machine learning algorithms to construct predictive models to detect the presence of bacteria such as Escherichia Coli and Staphylococcus Aureus in raw meat and determine which model is best through accuracy and cross-validation. In this study, five machine learning algorithms are used which are K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Random Forest (RF), Naïve-Bayes Classifier (NB), and Artificial Neural Network (ANN). All models are implemented effectively each having an accuracy of 94.97%, 91.84%, 97.57%, 61.46%, and 66.84% respectively. A web application is created using the shiny package in R to attain a standalone application used to show the detected bacteria.