Alzheimer's disease has been extensively researched in order to gain a deeper understanding of the illness for better early predictability. There have already been many studies that use structural and textural features to identify patients with Alzheimer’s disease. This study used an image processing technique to separate patients into four categories based on the severity of their diagnosis: mild-to-moderate, low, moderate, and severe. One method for identifying Alzheimer's traits is also examined in this study. Multi-class prediction 3 important for developing computational aids for differential diagnosis. The extraction methods for Alzheimer's disease are based on research comparing Naive Bayes (NB), Decision Tree, K – Nearest Neighbor, Gray Level Co-occurrence Matrix (GLCM), Principal Component Analysis, Local Binary Pattern (LBP), or Gray Level Co-occurrence Matrix (GLCM). The LBP and K‑nearest neighbor classification technique performed best on medical images with an accuracy of 87.96 percent.

PCA, GLCM, LBP, NB, DT, K-NN, MRI