Detection of drug dosage for cardiac patients using improved fuzzy convolutional neural network

Cardiovascular diseases are one of the most important causes of death in the whole world. If this disease can be identified quickly and accurately, it can be treated. It is very important to use different data to diagnose various diseases, especially cardiovascular diseases using data mining. Data mining on these data causes useful patterns to be obtained from these data. The biggest challenge in this field is the large amount of data. Therefore, using the right technique to get the right patterns with the help of data mining will help a lot.

In this article, a method based on the fuzzy optimal convolutional neural network that extracts the optimal features will be presented in order to prescribe the drug dosage for heart patients in addition to diagnosing heart disease. The weights of the used neural network will be optimized after training with the help of PSO. In this article, two types of data are used, which include the characteristics of patients and ECG of patients. A feature called drug dosage is also suggested, which prescribes the drug dosage based on the diagnosis of the severity of the heart disease.

The performance of the proposed method compared with other intelligent methods and it was saw that the proposed method performs better in the detection with an accuracy of 95.1% using the MIT-BIH and UCI databases simultaneously. It was also saw that training the network with a sufficient number of data makes all methods perform better. In addition, it was saw that if feature extraction such as wavelet transform, Fourier transform and PCA is performed, there will be an effective impact on the accuracy of the proposed method.

Originality/Value:

 to diagnose cardiovascular disease, a fuzzy convolutional neural network is used, whose network weights are optimized using PSO. In addition to the diagnosis of cardiovascular disease, the proposed method has the ability to prescribe the drug dosage.