Brain Stroke Risk Prediction using Hybrid Bayesian Boosting +Distance-based Interpolative Projections Interventions into Healthcare

Brain stroke is a primary cause of disability and death, necessitating an early and correct diagnosis for successful treatments. This paper presents a hybrid Bayesian Boosting with Distance-Based Interpolative Projections (BB-DIP) framework for classifying and predicting brain strokes using CT imaging dataset. The proposed BB-DIP model combines Bayesian boosting, a probabilistic ensemble learning approach, with distance-based interpolative projections (DIP), which optimize feature transformations by retaining important non-linear relations in a low-dimensional space. We compare the performance of BB-DIP with standard machine learning models like K-Nearest Neighbours (KNN) and Naive Bayes (NB). In terms of accuracy, precision, recall, and ROC-AUC, complete tests show that BB-DIP works better than the baseline models. The BB-DIP framework achieves highest accuracy of 94.8% and an AUC of 0.96, showing that it can deal with complicated features and make classification more reliable. This hybrid technique offers a viable alternative for improving stroke prediction and diagnosis by using medical imaging data.