Internet of medical things-based multitiered and hybrid architectural framework for effective heart disease prediction model

Cardiovascular and chronic respiratory diseases are becoming a global threat. The alarming fact about Cardio-vascular diseases (CVD) is that rural areas are recording a high number of heart failures compared to urban areas. The proposed research work intends to create an effective medical setup in country-wide through Internet of Medical Things (IoMT)-based multitiered and hybrid architectural framework and Machine Learning (ML) model. The proposed architecture collects the medical data using the IoMT devices. During data preprocessing, categorical data are encoded using one-hot mechanism and is normalized using standard scalar technique. In addition, feature selection mechanisms are applied to choose the best significant features using hybrid sequence of Filter-Wrapper-Ensemble approach. To reduce the diagnostic barriers in health prediction model, an efficient method of Bayesian optimization for hyperparameter tuning and k-fold cross-validation resampling technique is applied to evaluate various conventional and ensemble ML techniques. The prediction results show the high accuracy of 86.47% and 85.81% using Logistic Regression and Extreme Gradient Boosting (XGBoosting) algorithm, respectively. In particular, the accuracy obtained using decision tree for the proposed approach is higher (80.53%) than that of existing approach (73.22%).