Semi-Supervised Adaptive Neuro Fuzzy Inference System (SSANFIS) for Effective Classification of Alzheimer's Disease

Alzheimer's disease (AD) is the most popular dementia in elderly people worldwide. It affects memory of the patient. The early detection of Alzheimer’s disease still a challenge because of the estimation of the scans depends on manual directing and visual reading. To overcome this issue, Semi-Supervised Adaptive Neuro Fuzzy Inference System (SSANFIS) is introduced for effective classification of alzheimer's disease. In this work, an Improved Artificial Bee Colony (IABC) algorithm is used for preprocessing step which provides higher classification performance. It removes noises from the images based on the employed bees, onlooker bees and scout bees calculation. Then using the Adaptive Median Filter (AMF) enhances the quality of the images by comparing each pixel in the image to its surrounding neighbor pixels. The Hybrid Wavelet Transform (HWT) is enhanced using Haar with Walsh wavelet transform; it is used to extract the features from the MRI images. Unsupervised Fuzzy C Means (USFCM) is applied for selecting the important features from the extracted features. The model is learned by using Semi-Supervised Adaptive Neuro Fuzzy Inference System (SSANFIS) approach. It selects optimal fuzzy rules based on the higher frequency rules which are used to increase the accurate classification results which provide higher results than other methods