Clinical Risk Prediction of Acute Coronary Syndrome Using Deep Learning Approach

Abstract

Acute coronary syndrome (ACS) is a severe cardiovascular disease which is leading to death, and it is serious long-term disability globally. Prediction of ACS is important for diagnosing earlier to treat it. Previous ACS models based on small set of risk factors and predictive variables used to simplify the score calculation are numbered. This study has overcome the problem of existing system which developed stacked and regularized denoising predictive auto-encoder (SDAE) model to find clinical risks of ACS patients from huge collections of electronic health records (EHR) in biomedical engineering field. It determines patients at similar risk-level characteristics, low risk, high risk, and medium risk and preserves the results. This prediction approach is totally based on real-time dataset which we processed using SVM algorithm. Finally, SVM gives the accurate prediction of patients; if the prediction model shows the high risk, then doctor sends some precaution details to patients and his relatives for saving the life of patients. This approach is validated with more than 2000 real clinical dataset consisting of patient samples. The approach followed by us remains robust and also this model is very helpful for doctors and patients which can detect the cancer in an initial stage.