Liver Cirrhosis Stage Prediction Using Machine Learning: Multiclass Classification

Liver cirrhosis is a disease that affects a large population worldwide. Liver cirrhosis is further divided into four stages. This paper aims to predict the stage of liver cirrhosis of a patient using machine learning. It is a supervised learning problem of multiclass classification. Seven different algorithms were used for this purpose, and their performance was analyzed and compared in order to find the best approach. Different scaling and feature selection strategies were used in order to study their effect on the performance of various algorithms. It was found that an ANN-based approach achieves the best performance for this particular problem. A feature selection approach based on random forest and mutual information (RF + MI) was proposed in this paper, and its performance was compared with the standard Random Forest (RF) method for feature selection in classification problems. Experimental results demonstrated that the RF + MI approach shows minor improvement in comparison with random Forest (RF) for feature selection.