Classification Framework for Autoimmune Liver Disease using Machine Learning and Deep Learning Techniques
E. Ben George, G. Jeba Rosline, Abraham Varghese, Rajesh D. Gnana
Abstract
Autoimmune diseases affect people of all ages, genders, and ethnic backgrounds, making it one of the most widespread health issues we face today. These disorders occur when the immune system mistakenly targets healthy cells within the body and causing a wide range of potentially life-threatening diseases. There are over 80 recognized autoimmune disorders, including type-1 diabetes, rheumatoid arthritis, Graves’ disease, alopecia, celiac disease, and autoimmune liver disease(AiLD). Autoimmune liver disease is one such chronic disorder characterized by liver inflammation and damage. If not detected on time and untreated within the given time, it might lead to liver cirrhosis or liver cancer. Identifying and treating autoimmune liver disease is a complex and time consuming process which involves symptom analysis, laboratory tests, and clinical assessments. Detecting the disease can be more challenging because many symptoms of autoimmune liver diseases overlap with those of other autoimmune conditions. As a result, there is a need for an optimized, accurate, efficient, and cost-effective methodology to identify the disease without extensive and invasive procedures such as organ transplants. This paper introduces an innovative approach that employs the power of supervised machine learning and deep learning techniques to predict the onset of autoimmune liver diseases in advance. The machine learning model is constructed using a comprehensive dataset comprising the complete medical history of patients, physical examination records, histopathology and genetic data. Along with the machine learning model, the deep learning model is also developed for liver MRI images of various patients to make accurate prediction. This system provides a second opinion for medical professionals that may avoid the need for extensive and invasive medical procedures, including biopsies, and ultimately improve patient care and outcomes.