Abstract

Info

Title:: Classification and prediction of Alzheimer’s disease based on machine learning

Author(s):: LI Cai1; FAN Zhao2; 1. School of Basic Medicine, Shanxi Medical University, Taiyuan 030001, China; 2. Translational Medicine Research Center of Shanxi Medical University, Taiyuan 030001, China

Keywords:: Keywords: Alzheimer’s disease; machine learning; L1-regularized Logistic regression; classification and prediction

Abstract:: Abstract: Objective To classify and predict Alzheimer’s disease by machine learning method with brain structural magnetic resonance imaging (sMRI), age, gender, years of education and MMSE score as features. Methods After feature selection, L1-regularized Logistic regression, L1-regularized support vector machine and gradient boosting decision tree were used to classify and predict the brain sMRI data. After the optimal model was determined, the model was optimized by several features including age, gender, years of education and MMSE scores, and the performance of the optimized model was evaluated by 10-fold cross- validation. Results L1-regularized Logistic regression had the best classification performance. The prediction accuracy of the model optimized by the features of age, gender, years of education and MMSE scores was increased by 0.89% to 11.42%. Conclusion The feature set of sMRI, age, gender, years of education and MMSE scores in L1-regularized Logistic regression model has a better performance on the classification of Alzheimer’s disease, which can be used as the basis for the auxiliary diagnosis of Alzheimer’s disease.