Predicting Alzheimer’s Disease Diagnosis Using Cognitive, Lifestyle, and Medical Factors: A Logistic Regression Approach

Authors

• Lilly Fattizzi, University of Central FloridaFollow
• Josephine Schumacher, University of Central FloridaFollow

Streaming Media

Files

Course Code

STA

Course Number

4504

Faculty/Instructor

Professor Nathaniel Simone

Faculty/Instructor Email

nathaniel.simone@ucf.edu

About the Author

Lilly Fattizzi is an undergraduate student at the University of Central Florida, majoring in Psychology with a minor in Statistics. Josephine Shumacher is an undergraduate student at the University of Central Florida, majoring in statistics. This project was completed as part of coursework in categorical data analysis and reflects our interest in applying statistical methods to real-world health outcomes, particularly in the context of neurological and cognitive disorders.

I would like to thank Professor Nathaniel Simone for his guidance throughout this project. Additionally, we acknowledge the creator of the dataset used in this study for making it publicly available for educational and research purposes.

Abstract, Summary, or Creative Statement

This study examines how well cognitive performance and lifestyle/health factors predict Alzheimer’s disease diagnosis using a logistic regression framework. The dataset, obtained from Kaggle, consists of 2,149 individuals and includes demographic, cognitive, and medical variables. Key predictors of interest include Mini-Mental State Examination (MMSE) score, memory complaints, body mass index (BMI), age, depression, and hypertension. Exploratory data analysis revealed that individuals diagnosed with Alzheimer’s disease had substantially lower MMSE scores and higher rates of memory complaints compared to those without a diagnosis. A logistic regression model was fit, and stepwise selection based on Akaike Information Criterion (AIC) was used to identify the most important predictors, with the final model including MMSE, memory complaints, and hypertension. Results indicated that higher MMSE scores were associated with decreased odds of Alzheimer’s disease, while memory complaints significantly increased the likelihood of diagnosis. Model performance was evaluated using a receiver operating characteristic (ROC) curve, yielding an area under the curve (AUC) of 0.7189, indicating moderate predictive ability. Overall, cognitive variables were the strongest predictors of Alzheimer’s disease diagnosis, underscoring the importance of cognitive screening measures for early detection.

Keywords

Alzheimer’s disease; Logistic regression; Cognitive decline; MMSE; Memory complaints; Predictive modeling; Health data; ROC curve; Categorical Data Analysis;

Recommended Citation

Fattizzi, Lilly and Schumacher, Josephine, "Predicting Alzheimer’s Disease Diagnosis Using Cognitive, Lifestyle, and Medical Factors: A Logistic Regression Approach" (2026). High Impact Practices Student Showcase Spring 2026. 9.
https://stars.library.ucf.edu/hip-2026spring/9