Diagnosis of Liver Diseases using Machine Learning

Abstract

Computer aided diagnosis on medical domain such as breast cancer detection, brain tumor, liver disease, etc plays a major rile to go for smart hospitality. Liver diseases are responsible for more than 2.4% of annual deaths in India. Early detection of liver conditions is challenging due to the mildness of initial symptoms, which often become noticeable only at advanced stages. This paper seeks to enhance liver disease diagnosis by investigating two methods for identifying patient parameters. Due to liver diseases many peoples across the world lost their lives and its death rate can be reduced only by diagnosing disease on time but the main problem is LIVER will not show any symptoms for earlier damage. So in this paper is applying two methods to predict liver disease. Method1) in this method author is using INDIAN LIVER dataset to train various machine learning algorithms such as SVM, ANN and multilayer perceptron and this trained model will be applied on new patients TEST data to predict liver is normal or not but student ask us to implement Logistic Regression, Naïve Bayes and then compare its performance with SVM so we are using student suggested algorithms. Method2) in this method author is training ANN and CNN with gene MRNA images dataset and then training with CNN and ANN to predict whether liver disease inheriting in genes from ancestors. Student also asking to used liver images and then train with CNN and ANN; we are using LIVER ULTRA SOUND SCAN IMAGES