ABSTRACT
In citizens, Alzheimer's disease is Disease disease (AD) became very common. The early detection of such this this disease may save many lives because during this disease, people are not able to remember things because of damage to brain cells, which puts their lives in danger. In the last few decades, machine learning algorithms have played a very important role in the diagnosis of diseases in the medical field. In this chapter, we used magnetic resonance imaging (MRI) of the Alzheimer Disease Neuroimaging Initiative (ADNI) public dataset for our experiments. The dataset contains images of four classes, Alzheimer's diseasee (AD), mild cognitive impairment (MCI), late mild cognitive impairment (LMCI), and healthy persons, which shows the different stages of the disease. Then, we performed experiments in two parts. In the first part, we applied 11 basic machine learning algorithms algorithms, named random forest classification, extra trees classification, decision trees classification, XgBoost classification, Adaboost, K-nearest neighbor classification, logistic regression classification, NuSVC (Nu- support vector classifier) classification, linear SVC (support vector classifier) classification, gradient boosting classification, and i-layer perceptron neural network. The highest accuracy, 85.7%, we got for the random forest algorithm. Thereafter, one of the most effective deep learning models, and DenseNet was applied. We have considered DenseNet121, DenseNet169, and DenseNet201 in further experiments because this deep learning model can not only pick the features from input but also can propagate to the next layer, and the results show an approximately 9% increase in accuracy of classification then from the traditional machine learnings, which justifies the usefulness of the proposed DenseNet models.
