AI-Driven Forecasting Mechanism for Cardiovascular Diseases: A Hybrid Approach using MLP and K-NN Models
Koushik Reddy Chaganti, Pokuri Venkataradha Krishnamurty, Addala Hemantha Kumar, G. Sudha Gowd, Chinnala Balakrishna, P. Naresh
Abstract
Cardiovascular disease (CVD) results in malfunctioning of the heart and also the related blood vessels and regularly primes to death or corporal paralysis. Initial and automated discovery of CVD is crucial for saving many of the common humanoid lives. While several Investigations have been carried out to achieve this goal, and there is still room for development in terms of performance and reliability. This new work provides one more step in this area, since it utilizes two reliable ML algorithms, “multi-layer perceptron (MLP)” and “K-nearest neighbour (K-NN)“to detect CVD using publicly available data from the University of Engineering California Irvine source. The models’ concerts are enhanced optimally and by eradicating the possible outliers and common existing attributes with values that are treated as null. They performed general Experimental and effective findings show that the MLP model outperforms the KNN model in terms of reporting accuracy ($82.47 \%$) and area below the curve ($86.41 \%$). As a result, the proposed MLP model is recommended for automated CVD identification. Furthermore, this approach may be used to identify new illnesses, and the suggested model’s routine can be tested against other standard datasets.