Liver Disease Prediction Using Machine Learning Algorithms

Abstract

In Human beings, Liver is the most primary part of the body that performs many functions including the production of Bile, excretion of bile and bilirubin, metabolism of proteins and carbohydrates, activation of Enzymes, Storing glycogen, vitamins, and minerals, plasma proteins synthesis and clotting factors. The liver easily gets affected due to intake of alcohol, pain killer tablets, food habits, and includes plenty of wired practices. Currently, the liver related diseases are identified by analyzing liver function blood test reports and scan reports. It takes more time as well as expensive. While employing different data mining algorithms to ease this process it is possible to reduce the time for diagnosing the liver disease. When more data are used, the prediction will be more accurate. To avoid the local storage scarcity experienced in many healthcare centres, cloud storage is used. As the documents generated are voluminous in size in health care centres cloud storage would be an appropriate choice. This article discusses different data mining algorithms like K-Nearest Neighbor (KNN), Decision Tree (DT) and Adaptive Neuro-Fuzzy Inference System (ANFIS) that are used to provide a decision support model that could help the physician in predicting the liver disease from the dataset. The performance of each algorithm is evaluated with respect to accuracy, sensitivity, precision and specificity. A survey on the efficiency of these algorithms is presented.

Topics & Concepts