A Meticulous Presaging of Heart Disease Optimized by Boruta feature Selection and RFE over Gradient Boosting

In the medical field, healthcare is a vast domain by which the use of data science the meaningful information transformation is a necessity of healthcare. Due to the advancement in technology with the help of machine learning easily predict the disease. Feature selection comes under the category of machine learning.it is very prominent with variables and features especially when used in dataset. Feature selection improves accuracy and eliminates irrelevant features to achieve classification performance. Random forest and gradient boosting is quite emerged and useful algorithm by which the higher number of variables of feature selection can handle.it is a major issue in feature selection. The aim of this work is the popular dataset Cleveland (which has 303 instances) used with a higher number of parameters. Feature selection Boruta and RFE are used to get the best results and accuracy performance as compare to other existing works. To predict the possibilities patients have an HD firstly analyze the dataset and observations consider for 14 features. This examination breaks down the characteristics impact on the result of coronary illness. The ML approaches used for analysis Logistic Regression, Support Vector Machines (SVM), and Random Forest, Gradient Boosting, etc. the approaches and after-effects of the machine learning cross-validation and implementation. This present work objective was to focus on the relevant features of patients. GB, Boruta, and RFE are used to select the most relevant features. Boruta is based on the wrapper method of feature selection. This paper aims to find correlated feature accuracy and other combined with machine learning classifiers which help to find the robust prediction results. This proposed model achieved accuracy for GB classifier when implementing CF gives accuracy 83.82% and the for hybrid method (RF+GB+RFE+ Boruta) accuracy is 96.74% and for FIB 90.69% and F-score 96.74 %.