A Bayesian Method for Gastric Disease Diagnosis Estimation Among USU Mathematics Students

Abstract

Gastric diseases are frequently experienced by university students, especially Mathematics students at Universitas Sumatera Utara, due to demanding academic activities, irregular dietary patterns, and academic pressure. These conditions often lead to symptoms such as epigastric discomfort, nausea, vomiting, abdominal bloating, and early satiety. Since similar symptoms may arise from different gastric disorders, including dyspepsia, gastroesophageal reflux disease (GERD), gastritis, and Helicobacter pylori infection, an analytical approach is required to estimate the most likely diagnosis based on observed symptoms. This study employs Bayes’ Theorem to estimate gastric disease diagnoses using symptom data obtained from questionnaires distributed to students. Prior probabilities were determined based on the proportion of each disease within the observed sample, while conditional probabilities were calculated from the frequency of symptoms associated with each disease category.The analysis was carried out through manual calculations and supported by Microsoft Excel. The findings reveal that GERD exhibits the highest posterior probability in a specific test case, reaching 95%, indicating it as the most probable diagnosis. These results suggest that Bayes’ Theorem can be effectively utilized as a supporting method for estimating early-stage gastric disease diagnoses among university students.