Iranian Journal of Biomedical Engineering (IJBME)
نوع مقاله : مقاله کامل پژوهشی
نویسندگان
1 Department of Computer Engineering, Ayandegan Institute of Higher Education, Tonkabon, Iran.
2 گروه انفورماتیک سلامت، پژوهشکده تروما، دانشگاه علوم پزشکی گیلان، رشت، ایران
3 Department of Computer Engineering, Ayandegan Institute of Higher Education, Tankabon, Iran
کلیدواژهها
موضوعات
عنوان مقاله English
نویسندگان English
Heart disease is one of the leading causes of death in the world. If the current trend continues, 23.6 million people will die from heart disease by 2030; therefore, its accurate prediction can help reduce the mortality rate. Previous models have mainly used traditional machine learning algorithms, which have limitations in accuracy and generalizability. In this study, a new hybrid model based on Adabost and Convolutional Neural Network (CNN) is presented, which makes significant improvements in heart disease prediction by utilizing the power of deep learning and reinforcement algorithms. The dataset used includes data from Cleveland heart patients with 303 samples and 14 main features. Due to the imbalance of the data, the data of the two classes were first balanced using the random oversampling technique, so that each class included 242 samples. After balancing, the dataset was divided into two training and testing parts with a ratio of 80 to 20. The proposed model was compared with five popular algorithms including support vector machine, decision tree, random forest, logistic regression and naive Bayes. The results showed that our hybrid model outperformed the other methods with an accuracy of 87%. This study demonstrates that the combination of CNN adabus, along with data balancing, can be used as an efficient method for predicting heart disease.
کلیدواژهها English
