Automated Liver Disease Diagnosis using Machine Learning Techniques

Abstract

Liver diseases are among the most deadly  and panic conditions of health sector in entire world, because it is predicted that they will worsen due to a number of factors, including an increase in alcohol consumption, deteriorating global pollution from heavy industrialization and global warming, toxic gas exhaustion, contaminated water, food, and drugs, and most importantly, poor lifestyle choices. These factors all contribute to an ongoing rise in the diagnosis of liver anomalies in patients. Sometimes it is very difficult to identify the cause and symptoms by doctors. We are applying ML algorithms for the prediction of liver diseases. In order to develop  forecasting system for the early identification of liver disorder, the patients liver datasets are examined to  develop different forecasting system for the pre-symptomatic of liver cancer. Machine learning technique have the potential to significantly  improve the prediction  of accuracy. It has been observed that  machine learning is improving our basic understanding of illness progression. Application of machine learning algorithm plays a crucial role in analysing the liver datasets. The major goal of the investigation is to apply various ML algorithm for the prediction of liver diseases by comparative analysis of various machine learning models indicates the best method.  The prime purpose of this  work is for analysing the result  of LR, NB and Random forest computation in Liver disorder dataset. And Datasets are collected from the UCI database of Indian liver patient records. The suggested technique is executed using Python and outcomes are analysed ont the basis of accuracy, recall and precision. It is found that 0.75 of accuracy in Linear Regression, 0.74 of accuracy in Random Forest, 0.69 in Decision Tree, 0.64 of accuracy in Supporting voting machine, 0.62  of accuracy in knn and 0.53 of accuracy in Naïve Bayer. From above analysis Linear Regression having highest accuracy in compare to others. Our main focus on prediction of liver disease using clinical data.