Prediction of Liver Disease using Gradient Boost Machine Learning Techniques with Feature Scaling

Lifestyle diseases have become common these days and a sedentary way of life has paved the way for a range of syndromes and unknown diseases. Identification or diagnosis of the disease at an early stage is most crucial. This greatly helps in the prevention of the disease at an early stage with minimal medications. Traditional methods involve physical examination and lab results. Identification of the Liver disease at an early stage is very difficult as the symptoms of the diseases are visible only at a later stage of the disease. The Application of Machine learning models would help in the early diagnosis of the disease and hence facilitates in identifying crucial factors that lead to liver damage. In this paper, we propose a method of feature reduction using Recursive Feature Elimination and applying the Machine learning boosting algorithms to enhance the prediction accuracy. Basic machine learning models were applied to the dataset, where Logistic regression and Multi-Layer Perceptron had higher prediction accuracies with reduced features. Boosting algorithms like CatBoost, LGBM Classifier, XGBoost and Gradient Boost were applied to the dataset. The impact of feature reduction was investigated on the Gradient boosting machine learning algorithms.