Early Detection Of Liver Disease Using Machine Learning Models

Liver diseases are one of the major global health problems, and better patient outcomes and treatment are reliant on early diagnosis. This research considers a dataset that was obtained from the North East of Andhra Pradesh, India, and contained data from 416 patients with liver disease and 167 individuals without liver disease. Many clinical and demographic factors, such as age, gender, levels of bilirubin, alkaline phosphatase, and liver enzyme markers, are part of the collection. The principal aim of this project is to create a model that, based on patient information, has the ability to predict liver disease. In order to gain a deeper insight into the dataset, exploratory data analysis is performed, showing key trends and patterns. Missing values, especially in the "Albumin and Globulin Ratio" column, are managed with appropriate imputation techniques. Statistical analysis also offers important insights into the variability and central tendencies of the dataset. Visualizations, including histograms and count plots, assist in portraying the distribution of variables such as age, gender, and the target variable used to denote liver disease. Predictive models will be developed by applying machine learning methods like logistic regression, decision trees, and support vector machines. Measures like accuracy, sensitivity, and specificity will be utilized to train and test these models in order to measure their performance. The project has the ability to help healthcare professionals identify those at risk of liver disease early on, thus enabling more preventive healthcare interventions and enhancing patient outcomes through evidence- based insights.