Logistic random forest boosting technique for Alzheimer's diagnosis

Area of Science:

• Neuroscience and computational biology
• Medical imaging analysis
• Machine learning applications in healthcare

Background:

• Alzheimer's disease (AD) is a prevalent neurodegenerative disorder causing cognitive impairment, with no current cure.
• Effective management of AD symptoms requires timely intervention tailored to disease stage.
• Early detection and classification of AD are crucial for optimizing patient care and quality of life.

Purpose of the Study:

• To develop and evaluate a machine learning model for early identification and classification of Alzheimer's disease stages.
• To differentiate between Alzheimer's disease (AD), Cognitively Normal (CN), and Late Mild Cognitive Impairment (LMCI) using neuroimaging data.
• To assess the performance of an ensemble model against various individual and ensemble machine learning algorithms.

Main Methods:

• Utilized the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset for model training and validation.
• Developed an ensemble model named Logistic Random Forest Boosting (LRFB), combining Logistic Regression (LR), Random Forest (RF), and Gradient Boost (GB).
• Compared the LRFB model's performance against multiple established machine learning algorithms including k-NN, MLP, SVM, AdaBoost, Naïve Bayes, XGBoost, and Decision Tree.

Main Results:

• The proposed LRFB ensemble model demonstrated superior performance across key metrics: Accuracy (Acc), Recall (Rec), Precision (Prec), and F1-Score (FS).
• LRFB significantly outperformed individual models like LR, RF, GB, and other ensemble methods in classifying AD, CN, and LMCI.
• The study highlights the effectiveness of ensemble methods in enhancing diagnostic accuracy for neurodegenerative diseases.

Conclusions:

• The LRFB model shows significant promise for accurate and early detection of Alzheimer's disease stages.
• This machine learning approach can aid clinicians in timely diagnosis and personalized treatment strategies for AD patients.
• Further research and validation are warranted to integrate this model into clinical diagnostic workflows.