Prediction and Influencing Factor Analysis of Alzheimer’s Disease Based on Machine Learning Algorithms - Research and Model Evaluation

Aug 4, 2026

Current research on Alzheimer’s disease remains incomplete and faces challenges in achieving precise, scalable diagnosis. This study investigates underlying causes by leveraging machine learning algorithms and predictive models. Three models—Random Forest, XGBoost, and Support Vector Machine—are compared using grid search with five-fold cross-validation, optimizing parameters and constructing final predictors. Model quality is assessed with accuracy, recall, and weighted F1 score, where XGBoost achieves the strongest results and key predictive features are identified.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Abstract

Introduction

Alzheimer’s disease challenges

Why machine learning

Related Work

Disease prediction with ML models

ML approaches in Alzheimer’s research

Methods and Evaluation

Model comparison (RF, XGBoost, SVM)

Parameter tuning and cross-validation

Evaluation metrics and results

Prediction and Influencing Factor Analysis of Alzheimer’S Disease Based on Machine Learning Algorithms Ruoyao Ge 1 􀀍, Mingzhen Zhou2, Ziqi Zhou3 1College of Statistics of Capital University of Economics and Business, Beijing, China 2Shenzhen Senior High School, Shenzhen, China 3Beijing Guangqumen Middle School, Beijing, China Abstract. Current research on Alzheimer"s disease remains incomplete and faces significant challenges. This study aims to investigate the underlying causes of Alzheimer"s disease by leveraging machine learning algorithmsand predictive models. The performance of three machine learning models including Random Forest (RF), XGBoost, and Support Vector Machine (SVM) were compared in predicting Alzheimer's disease. By employing grid search and five-fold cross-validation, we identified the optimal parameters for each model and constructed predictive models based on these parameters. For model evaluation, we used accuracy, recall, and the weighted average F1 score as metrics. The results demonstrated that the XGBoost model performed better than the others, achieving accuracy, recall, and F1 scores of 95.59%, 95.58%, and 95.56%, respectively. In contrast, the SVM model showed slightly lower performance, with all three metrics hovering around 89% . Therefore, the XGBoost model is considered the most suitable for Alzheimer's disease prediction. Additionally, the study identified key predictive features, including FunctionalAssessment, ADL, MMSE, MemoryComplaints, and BehavioralProblems. Future research could investigate ensemble learning techniques to further improve the model's predictive performance 1 Introduction Alzheimer's Disease (AD) is an irreversible neurodegenerative disorder of the central nervous system. Currently, with the continuous increase in the aging population globally, the number of Alzheimer’s patients is also on the rise. However, Traditional diagnostic and research methods for Alzheimer ’s disease often rely on subjective clinical assessments and limited biomarkers, which may lack the precision and scalability needed to address the disease's complexity. Additionally, these methods struggle to integrate and analyze the vast, multidimensional data required to uncover subtle patterns and interactions critical for early detection and personalized treatment strategies. 􀀍[32022010129@cueb.edu.cn](32022010129@cueb.edu.cn) © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 ([https://creativecommons.org/licenses/by/4.0/](https://creativecommons.org/licenses/by/4.0/)). Machine learning algorithms can handle multiple variables simultaneously and construct complex models, which helps to reveal the non-linear relationships among these influencing factors in Alzheimer ’s disease. Secondly, machine learning can comprehensively analyze multiple sets of medical data to help determine whether patients are at risk of Alzheimer ’s disease. This enables patients to receive early intervention. In addition, prediction models based on machine learning can assist public health departments in identifying high-risk groups for Alzheimer ’s disease, thereby carrying out targeted prevention efforts. Currently, numerous scholars are conducting research on multiple machine learning models in the field of disease prediction and analysis. Vishwanatha et al. deeply explored the use of machine learning in the domain of Chronic Kidney Disease (CKD) . They employed the Support Vector Machine (SVM) model to develop and validate a predictive model using CKD data features, aiming to estimate the likelihood of future renal replacement therapy. It was confirmed that the model is feasible and has a relatively high accuracy [1] . Liang et al. collected data from 9,171 patients to construct multiple machine learning models to predict major adverse cardiovascular events during the perioperative period. It was shown that a

FAQs

Which machine learning models were compared for predicting Alzheimer’s disease?

Random Forest (RF), XGBoost, and Support Vector Machine (SVM) were compared for Alzheimer’s disease prediction.

How were the model parameters tuned and validated?

Grid search combined with five-fold cross-validation was used to determine the optimal parameters for each model.

Why is XGBoost considered the most suitable model in this study?

XGBoost achieved the highest performance across accuracy, recall, and weighted F1 score, outperforming RF and SVM in the reported results.

Aug 4, 2026

Upload by Genevieve