Comparative Analysis of Using Artificial Intelligence (AI) for Diagnosis and Treatment of Tuberculosis and Diabetes
Authors: John Christian Co, Curt Mitchell Lameseria, Raymond Paiton
Abstract
The world’s epidemic and population killer are diabetes and tuberculosis. A lot of people are poorly diagnosed and untreated prior to death. These diseases cause patients slow weakening of the immune system. As the quote states, “prevention is better than cure”, this is what Artificial Intelligence (AI) is to offer. AI is developed in so many applications; among them is AI in medicine and healthcare. AI compared to conventional methods of disease detection is a huge help for patients to cope up early with their sickness. With AI providing early disease detection to tuberculosis and diabetes, patients with these illnesses can prevent severe conditions of it, better yet – avoidance. Differentiating the characteristics of these illnesses, it also has its own conventional methods of testing. Tuberculosis uses sputum sampling, skin test, blood test, and Xray while diabetes is diagnosed via fasting blood sugar. AI algorithms such as Fuzzy, Support Vector Regression, Naive Bayes Classifier, Decision trees, Genetic, AdaboostM1, Random committee classifier, and many more are the ones used for the study. These algorithms’ detection accuracy rates on the above-mentioned diseases were the factor of comparison in concluding and recommending for the future directives of these developments. Among those machine learning techniques, Support Vector Machine with AIRS exhibited a 92% to 100% (near-perfect-to-perfect) range in performance for Tuberculosis vs to a 83.0% accuracy in the training set and 76.9% in an independent testing set for Diabetes. Thus, proving efficacy of AI disease detection on both, yet showing accuracy dominance on Tuberculosis detection. AI in medicine and healthcare is still just a tip of the iceberg. Healthcare and engineering combined results to help mankind. Hopefully, this will be the future to prolong the lives of the human race.