Cardiovascular disease classification based on a multi-classification integrated model

Cardiovascular disease (CVD) has now become the disease with the highest mortality worldwide and coronary artery disease (CAD) is the most common form of CVD. This paper makes effective use of patients' condition information to identify the risk factors of CVD and predict the disease according to these risk factors in order to guide the treatment and life of patients according to these factors, effectively reduce the probability of disease and ensure that patients can carry out timely treatment. In this paper, a novel method based on a new classifier, named multi-agent Adaboost (MA_ADA), has been proposed to diagnose CVD. The proposed method consists of four steps: pre-processing, feature extraction, feature selection and classification. In this method, feature extraction is performed by principal component analysis (PCA). Then a subset of extracted features is selected by the genetics algorithm (GA). This method also uses the novel MA_ADA classifier to diagnose CVD in patients. This method uses a dataset containing information on 303 cardiovascular surgical patients. During the experiments, a four-stage multi-classification study on the prediction of coronary heart disease was conducted. The results show that the prediction model proposed in this paper can effectively identify CVDs using different groups of risk factors, and the highest diagnosis accuracy is obtained when 45 features are used for diagnosis. The results also show that the MA_ADA algorithm could achieve an accuracy of 98.67% in diagnosis, which is at least 1% higher than the compared methods.