Advancing Vascular Dementia Diagnosis Through Salp Swarm Optimized Convolutional Neural Network

Vascular Dementia (VD) is a form of cognitive impairment seen in individuals, restricting from problems with blood vessels supplying the brain. These issues lead to strokes or reduced blood flow and gradually lead to the development of vascular dementia. Since it tends to progress slowly, diagnosis often comes after significant brain damage making prevention is crucial. Early identification of risk factors are vital to mitigate the risk of irreversible cognitive decline associated with vascular dementia. Accordingly, this research proposes an optimized neural network model for predicting vascular dementia at its earlier to improve the quality of life. The initial stage in this research involves preprocessing, where the input data is cleaned and normalized. These steps ensure that the vascular dementia dataset is cleaned of inconsistencies and prepared with normalized features for accurate analysis and modeling. Following preprocessing, relevant data features are selected using the chi-square approach, making the classification task easier. Finally, classification is performed utilizing a Convolutional Neural Network (CNN) and for fine-tuning of CNN parameters Salp Swarm Optimization (SSO) Algorithm is implemented. The execution of proposed work is processed using Python software and the outcome reveals superior performance in terms of accuracy and precision. Thereby the proposed method provides better performance with an accuracy of 94.5%, contrasted to existing methods in predicting VD.