Foodborne Disease Symptoms, Diagnostics, and Predictions Using Artificial Intelligence-Based Learning Approaches: A Systematic Review

Food-borne diseases have a high worldwide occurrence, substantially impacting public health and the social economy. Most food-borne diseases are contagious or poisonous and are caused by bacteria, viruses or chemicals that enter the body via contaminated food. The most prevalent harmful bacteria (Salmonella, Escherichia coli, Campylobacter, Clostridium and Listeria) and viruses (norovirus) may cause acute poisoning or chronic disorders such as cancer. Thus, the detection of pathogenic organisms is crucial for the safety of food. Artificial intelligence has recently been an effective technique for predicting pathogens spreading food-borne diseases. This study compares and contrasts the accuracy of many popular methods for making predictions about the pathogens in food-borne diseases, including decision trees, random forests, k-Nearest Neighbors, stochastic gradient descent and extremely randomized trees, along with an ensemble model incorporating all of these approaches. In addition, principal component analysis and scaling methods were used to normalize and rescale the values of the target variable in order to increase the prediction rate. The performance of classification systems has been examined using precision, accuracy, recall, F1-score and root mean square error (RMSE). The experimental results demonstrate that the suggested new ensemble model beat all other classifiers and achieved the average highest 97.26% accuracy, 0.22 RMSE value, 97.77% recall, 97.66% precision and 98.44% F1-Score. This research investigates the predictability of pathogens in food-borne diseases using ensemble learning techniques.