A New Cuckoo Algorithm Based Least Square Support Vector Machine for Classification of Alzheimer Disease from Human Brain MRI

Abstract

Alzheimer's disease (AD) is a sort of brain condition that leads to the loss of daily functioning. Early diagnosis and classification of Alzheimer's disease remain unexplored due to the rapid progression of Alzheimer's patients and the absence of effective diagnostic instruments. The detection of early morphological changes in the brain and early diagnosis of Alzheimer’s disease (AD) are important in the field of healthcare. It is possible to use high-resolution magnetic resonance imaging (MRI) to help diagnose and predict this disease. Machine learning techniques are widely used now for neuro-imaging based diagnosing. These strategies yield absolutely automatic clinical decisions, unbiased by variable radiological expertise. This analysis paper compares and evaluates the performance and effectiveness of standard Least Square Support Vector Machine (LSSVM) therewith of Cuckoo Search (CS) based LSSVM within the diagnosing of Alzheimer’s disease (AD) diagnosing. The manual interpretation of enormous volume of brain imaging and cognitive measures could result in incomplete diagnosing. The CS-LSSVM approach is trained with multiple biomarkers to facilitate effective, accurate classification that could be a demand of the hour. Wavelet based texture features and multiple biomarkers are fed as input to the classifier. CS-LSSVM yields ninety eight correct results and beat than fuzzy c-means classifier in terms of sensitivity, specificity and accuracy during this analysis.