Comparison of supervised techniques of artificial intelligence in the prediction of cardiovascular diseases.

Cardiovascular disease is the main cause of mortality world-wide, its early prediction and early diagnosis are fundamental for patients with this mortal illness. Cardiovascular disease is a real threat for the Health Systems worldwide, mainly because it has become the diagnosis that claim a significant number of lives around the world. Currently, there is a growing need from health entities to integrate the use of technology. Cardiovascular disease identification systems allow the identification of diseases associated with the heart, allowing the early identification of Cardiovascular Diseases (CVD) for an improvement in the quality of life of patients. According to the above, the predictive models of CVD have become a common research field, where the implementation of feature selection techniques and models based on artificial intelligence provide the possibility of identifying, in advance, the trend of patients who may suffer from a disease associated with the heart. Therefore, this paper proposes the use of feature selection techniques (Information Gain) with the variation of artificial intelligence techniques, such as neural networks (Som, Ghsom), decision rules (ID3, J48) and Bayesian networks (Bayes net, Naive Bayes) with the purpose of identifying the hybrid model for the identification of cardiovascular diseases. It was used the data set “Heart Cleveland Disease Data Set” with the same test environment for all the cases, in order to establish which of the mentioned techniques achieves the higher value of the accuracy metric when it comes to identify patients with heart disease. For the development of the tests, 10-fold Cross-Validation was used as a data classification method and 91.3% of the accuracy was obtained under the hybridization of the selection technique “information gain” with the training technique J48.

• Autores: Comas Gonzalez Z.; Mardini Bovea J.; Salcedo, D.; De la Hoz Franco, E.
• Tipo de recurso: Article of investigation
• Fecha de publicación: 2023
• Institución: Corporación Universidad de la Costa
• Repositorio: REDICUC - Repositorio CUC
• Idioma: eng
• OAI Identifier: oai:repositorio.cuc.edu.co:11323/13663
• Acceso en línea: https://hdl.handle.net/11323/13663; https://repositorio.cuc.edu.co/
• Palabra clave: Artificial intelligence; Cardiovascular disease; Multimodal physiological measures
• Rights: closedAccess
• License: Atribución-NoComercial-SinDerivadas 4.0 Internacional (CC BY-NC-ND 4.0)