Prediction of Liver Abnormality using Machine Learning

Cancer according to the American Society of Clinical Oncology journal was first described back in 1600 B.C. and has been prevalent ever since. The technological advancement in the field of medical sciences aided with advancement in the field of machine learning and deep learning has brought us to a situation today, where a subject can be informed of the dangers or the possibility of possessing an infected liver. In the prediction model, different enzymes were studied, and appropriate ratios were to determine the stability of the hepatocytes in the liver. The data was employed by different Machine Learning algorithms and based on their accuracy levels the final prediction has been made using the most appropriate algorithm. In an attempt to take the model to the next level, a few more algorithms were employed and explored the dataset even more. The results of each algorithm are compared using ROC graphs and ROC AUC SCORE to achieve a better model for this prediction model. Each algorithm is given by certain hyper-parameters which would increase the fitting nature more towards the best. The most important features calculated by each algorithm are mentioned and used accordingly to calculate the results.