A hybrid approach for medical images classification and segmentation to reduce complexity

The computational domain facilitates the performance of novel and innovative medical research and development tasks by providing support and computational power. This analysis method forecasts the future by analyzing the data we now have. The method may be divided into three primary phases: preprocessing, feature extraction, and classification. The research presented here aimed to improve the precision with which heart disease could be predicted across three distinct phases. The first step is thoroughly examining the databases kept at the UCI computer repository. In this study, we use the dataset’s five different algorithms, decision tree, Naive Bayes, random forest, KNN, and support vector machine, to compare their respective performances. In addition to age, the suggested revolutionary technique considers other characteristics such as pulse rate, cholesterol, and so on, which was not the case in earlier studies. In the past, age was the primary consideration in analysis and illness prediction. Compared to more traditional methods, improved prediction accuracy is achieved by modifying the study’s primitive properties. Third, this research introduced a novel hybrid classification model by fusing support vector machines and k-nearest neighbor classification techniques. A k-nearest neighbor classifier will do the heavy lifting to classify the data, while support vector machines will extract the dataset’s features. The accuracy rates for the various prediction methods decision tree, KNN, Naive Bayes, random forest, support vector machine, and proposed method range from 72.53% to 87.32% to 87.39% to 81.34%, respectively. The new technique decreases execution value by 5 % and increases accuracy by up to 8 %. The suggested model outperforms state-of-the-art approaches in terms of accuracy and implementation speed.