چكيده به لاتين
Today, with the rapid growth of databases in most medical centers and organizations, a huge amount of data is stored in databases, one of which is electrocardiogram tapes. In recent years, the tendency to search for hidden patterns in data to improve physicians' decisions has increased dramatically. Cardiovascular disease is one of the most common and dangerous diseases in the country and the world, so the medical community to minimize their effects, complications and costs seek programs for further investigation, prevention, early detection and effective treatment. It is. Recently, data mining methods have been used to achieve defined goals.
The present study has performed the automatic diagnosis and classification of heart diseases by electrocardiogram image processing and data mining method. These data were collected by the Institute of Endocrine Sciences and Metabolic Diseases of Tehran University of Medical Sciences. The data used in this study, which includes 1115 images of 12-lead ECG, were first processed with MATLAB software, then by determining the criteria for diagnosis and application of Minnesota code and genetic algorithms, fuzzy decision tree, simple Bayesian and method Bagging is based on Crisp's standard data mining process. Diseases were classified into 8 classes, of which 6 classes included ischemic diseases, 1 class included arrhythmia, and 1 class included normal ECG scans. In the end, this research, which is an epidemiological study, helps physicians to make decisions to prevent misdiagnosis (from patients' ECGs) and management decisions in the field of health.
