Volume 18, No. 6, 2021

Scaling Of Medical Disease Data Classification Based On A Hybrid Model Using Feature Optimization

Jayesh Mohanrao Sarwade , Harsh Mathur


The classification of medical disease data leverages the healthcare informatics system. The variant of medical disease data is very complex such as document, numerical and constant. Therefore, the classification algorithms face a problem of data transformation and are compromised with the accuracy and sensitivity of classification algorithms. The increased classification rate of medical disease data promotes telemedicine's automatic diagnosis of critical illness. This paper proposed a feature optimization-based hybrid model for medical disease data classification. The feature optimization of disease data applies particle swarm optimization, and optimal features constraints increase the classification rate of proposed algorithms. The hybrid model of the classification algorithm is the process of clustering and classification. The proposed algorithm uses K-means and support vector machines (KSVM). The KSVM algorithm boosts the process of classification of medical disease data. The hybrid algorithm has been simulated in MATLAB environments and tested with a series of reputed datasets such as diabetics, cancer, hepatitis, lymphography and other three datasets. The performance of the proposed algorithm compares SVM and CNN. The analysis of evaluation results suggests that the proposed algorithm increase 2-3% of accuracy and sensitivity of data classification.

Pages: 3681-3696

Keywords: Data classification, Healthcare, Machine learning, CNN, PSO hybrid Model.

Full Text