DEVELOPMENT OF AN AI-POWERED SYSTEM FOR LABORATORY IDENTIFICATION OF SELECTED ENTERIC BACTERIA

Abstract

Microorganisms, commonly referred to as microbes, are living organisms too small to be seen by the naked eye but observable with a microscope. This study presents the development of an AI-powered system for the laboratory identification of selected enteric bacteria. Leveraging machine learning algorithms and comprehensive datasets containing morphological and biochemical features, the system aims to enhance the accuracy and efficiency of bacterial identification. The process includes gathering relevant data and literature on laboratory identification methods for selected bacteria, compiling a comprehensive database of identification features, developing machine learning algorithms capable of identifying bacteria based on key features, and laboratory confirmation of selected isolates with the developed model. The dataset encompasses selected enteric bacteria: Pseudomonas species, Vibrio species, Escherichia species, Citrobacter species, Staphylococcus species, Salmonella species, Shigella species, Campylobacter species, Clostridium species, and Enterococcus species. Results from the analysis showed that the Support Vector Machine (SVM) and Random Forest models achieved the highest accuracy at 83%, while the XGBoost model reached 50%. Conversely, the Decision Tree and ANN models performed poorly with 16% accuracy each. These findings underscore the potential of AI-driven approaches, particularly SVM and Random Forest, to revolutionize bacterial identification, with significant implications for public health, research, and clinical practice.