Abstract:
Irritable Bowel Syndrome (IBS) is a complex and diverse non-communicable gastroin testinal disorder associated with overlapping symptoms, psychological disorders, various
co-morbidities, and phenotypes. Many theories have been put forward, including stan dardized Rome criteria but the exact risk locus of IBS is still uncertain. A careful medical
history is critical, particularly regarding possible comorbidities. The applications of data
science, are evolving in various domains of health and treatment. Applying techniques
from machine learning to medical data can bring vital, valuable, and effective achieve ments, enhancing clinicians’ knowledge and expediting the diagnosis process. This study
utilizes data analysis and unsupervised machine learning techniques on a multidimen sional clinical dataset to identify factors associated with IBS. It was observed that an
association exist between IBS and obesity or overweight status in females. By applying
clustering algorithms, associations were seen between IBS with medical conditions such
as ischemic heart disease, hypertension, atherosclerosis, hypercholesterolemia, and her nia. The evaluation process primarily concentrated on identifying the top twenty-five
comorbidities that exhibited the highest degree of association with IBS. These comor bidities were selected based on their significant contribution to the accurate clustering of
IBS, accounting for approximately 83% accuracy. Therefore, keeping highlighted comor bidities in view diagnostic confidence can increase and reduce the amount of investigation
in many patients with IBS.
