COMPARING CLASSIFICATION MODELS FOR PREDICTING LIVER DISEASES

There is a constant rise in the number of liver disease patients with the most common reason being alcohol overdose, intake of drugs and medicines. This research work focuses to classify the given data using Support Vector Machine and Artificial Neural Network and label the patients as having liver disease or not. SVMs are a non-probabilistic binary classifier which label the new data into one of the two belonging categories. Artificial Neural Networks are designed to solve problems like the human brain can. They consist of nodes (known as artificial neurons) which are arranged in layers and can transmit information in the form of real number. The objective of this research is to compare the performance of SVMs and ANNs to classify the patients and to improve the accuracy of ANN using k-fold cross validation and hyperparameter tuning. The experimental results show that ANNs outperform the SVMs. The results can help the doctors in diagnosing liver diseases.