Data Mining Based Heart Disease Prediction Using Hybrid Optimization Technique of Derived Features

Data Mining Based Heart Disease Prediction Using Hybrid Optimization Technique of Derived Features

연구 분야: Databases

학회: International Conference on Artificial Intelligence and Speech Technology

Data mining difficulties include improving forecasts, medical domain data from clinical data, and research observations. This section analyses and discusses medical data mining applications and research. Many medical treatment issues are investigated using data mining. One of the biggest challenges for medical service providers is providing high-quality treatment at low cost (such as hospitals and medical centres). Excellent service involves accurate patient evaluation and therapy management. Clinical choice is crucial to diagnosing a patient’s health. Data mining uses clinical data and scientific observations to improve medical data and projections. This article highlights several of the most critical issues facing this business for massive data collections and streams. The suggested study examines the dataset size, cost sensitivity, location, and heart disease prediction factor relevance. In sequence, the proposed activity incorporates dataset collection, feature extraction, feature selection, feature fusion, and data classification. Feature Selection is the most important phase in data mining using Hybrid Grey Wolf and Particle Swarm Optimization (HGWPSO). Classifying Cardiovascular Disease using an Enhanced Probabilistic Neural Network (EPNN). Because of this, we aim to highlight sensitivity vs. specificity in the graph despite the data’s high accuracy. Iterating over the standard classifiers yields the following results: the area under the ROC curve should be a maximum of 0.9063 and the accuracy should be 96.667%.