Construction and Validation of an Early Screening Model for Alzheimer's Disease based on Big Data

Abstract

Alzheimer’s disease (AD) is a serious neurodegenerative disease. The incidence of AD is increasing with global aging. This study utilized the Alzheimer's_disease_data dataset from the Kaggle website, which contains multi - index information on 2149 patients. After data preprocessing, Methods such as the t-test, analysis of variance, and chi-square test were used to analyze the relationships between numerical variables and categorical variables. Key indicators such as "smoking" and "family history of Alzheimer's disease" were screened out. Based on these results, four different early screening models, namely logistic regression, naive Bayes, K-nearest neighbors, and polynomial regression, were meticulously constructed. Subsequently, a series of indicators such as accuracy, sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC) were compared to validate and analyze each model. The confusion matrices were also analyzed to gain a deeper understanding of the performance of each model. This study provides a direction for the optimization of early screening models for Alzheimer's disease (AD), and is expected to provide more effective tools for clinical diagnosis and intervention.