Crohn’s and ulcerative colitis are two diseases classified under inflammatory bowel diseases that are frequently confused during the diagnostic process. These conditions are chronic diseases that significantly affect patients’ quality of life. Since they may present with similar clinical symptoms accurate and timely diagnosis is often difficult. Misdiagnosis may lead to inappropriate treatment and delay disease control, thereby imposing an additional burden on the healthcare system. The diagnosis of inflammatory bowel diseases is currently based on the combined evaluation of medical history, physical examination findings, biochemical tests, radiological assessments, endoscopic procedures, and histopathological examination of biopsy samples. However, this process is both invasive and limited in terms of time and cost. Therefore, there is an increasing need for less invasive and faster diagnostic approaches. The potential to establish a preliminary diagnosis using easily accessible data such as symptom information and blood tests may improve the efficiency of clinical processes. In this study, considering these challenges, machine learning approaches were applied to support the diagnostic process using logistic regression and random forest methods. The results indicate that both methods achieved successful performance in disease classification. However, the random forest method demonstrated better performance with higher accuracy.

Keywords

Inflammatory bowel diseases, Crohn’s disease, Ulcerative colitis, Machine learning, Classification