Volume - 13 | Issue-1
Volume - 13 | Issue-1
Volume - 13 | Issue-1
Volume - 13 | Issue-1
Volume - 13 | Issue-1
Cardiovascular disease (CVD) is a major cause of death, disability, and hospitalization worldwide. In this paper, a deep learning approach has been proposed to detect CVD by analyzing electronic health records (EHRs). This approach uses natural language processing (NLP) to extract features from unstructured text such as clinical notes and laboratory results, and then deploys deep learning models to predict the presence of CVD. Sentiment analysis has been combined with deep learning for CVD detection. This approach combines NLP methods with supervised machine learning (ML) algorithms to automatically extract subjective information from EHRs. The sentiment analysis model automatically extracts both the overall sentiment of a corpus and the sentiment expressed in individual sentences. Our results showed that the CNN and LSTM algorithms performed best in detecting the presence of cardiovascular diseases from medical documents, with an accuracy of 87.5% and 84.8%, respectively. The RNN and Naïve Bayes algorithms, on the other hand, had a lower accuracy of 76.1% and 68.7%, respectively. The extracted sentiment information is then used as another feature to further refine the deep learning model. This approach has been shown to outperform traditional ML models for CVD detection.