Machine Learning Based Prediction of Non- Alcoholic Fatty Liver Disease Using Clinical Parameters Kartikey Sharma, Satyam Chaurasiya, Mrs. Manshee Agarwal, Shiv Prakash Mishra,

Abstract: Non-alcoholic fatty liver disease (NAFLD) has recently emerged as one of the most rapidly spreading chronic liver diseases worldwide and a major public health problem. This disorder has been found to have a strong relationship with obesity, type 2 diabetes, metabolic syndrome, high blood pressure, and an unhealthy way of life. The prevalence of sedentary activities and high consumption of processed food is one of the reasons behind the rise in the prevalence rate of NAFLD cases in the world today. Often, people suffer from non-alcoholic fatty liver disease without exhibiting any symptoms in the initial stages of the disease. If this condition is not treated promptly, then it can develop into more severe health conditions such as liver inflammation, fibrosis, cirrhosis, liver failure, and even cancer of the liver. There are several methods of diagnosing NAFLD through conventional testing such as ultrasonography, CT scans, MRIs, blood tests, and liver biopsy.

The objective of this paper is to provide a machine learning-based web application that predicts the risk of developing NAFLD with the help of easily accessible health features including age, gender, height, and weight of the individuals. Body Mass Index (BMI) is dynamically generated on the basis of user inputs and included as a further input parameter in view of the strong association of BMI with obesity and metabolic syndrome. The Random Forest method is utilized in order to generate the NAFLD prediction model owing to high accuracy and stability in case of structured medical datasets. The prediction model will be able to detect the patterns related to NAFLD and classify the users as either low risk, medium risk, or high-risk individuals. Moreover, probability-based outputs have been considered to improve the comprehensibility of the results.

The experimental findings demonstrate that the system under development is able to generate fast, accurate, and efficient screening results. The system can be used to enhance healthcare awareness campaigns, conduct preventive diagnoses, and facilitate consultations by medical practitioners for high-risk individuals. The system may help medical practitioners in making decision when conducting an initial risk assessment. Further developments of the system can involve incorporating other health parameters like cholesterol, glucose, and liver enzyme levels, managing patient records securely, providing a multi-language environment, mobile compatibility, and cloud-based health service provision.

Keywords: NAFLD, Machine Learning, Random Forest, Flask, BMI, Liver Disease Prediction, Healthcare Analytics