Prediction of Alzheimer's in People with Coronavirus Using Machine Learning

Area of Science:

• Neurology
• Infectious Diseases
• Artificial Intelligence

Background:

• COVID-19 infection is linked to an increased risk of developing Alzheimer's disease.
• Cognitive impairment, or "oblivion," is a significant post-COVID-19 complication affecting many individuals globally.
• Early prediction of Alzheimer's disease in COVID-19 patients can mitigate the severity of neurological sequelae.

Purpose of the Study:

• To predict the onset of Alzheimer's disease in individuals previously infected with COVID-19.
• To evaluate the efficacy of machine learning algorithms for Alzheimer's disease prediction in the context of COVID-19.
• To identify the most accurate predictive model among Nave Bayes, Random Forest, and K-Nearest Neighbors (KNN).

Main Methods:

• A dataset was compiled from COVID-19 patients in Tehran Province, Iran, between October 2020 and September 2021.
• Three machine learning algorithms were employed: Nave Bayes, Random Forest, and KNN.
• Model performance was quantitatively assessed using Precision, Recall, Accuracy, and F1-score metrics.

Main Results:

• Both Nave Bayes and Random Forest algorithms achieved prediction accuracies exceeding 80%.
• The Random Forest algorithm demonstrated superior predictive accuracy compared to Nave Bayes and KNN.
• The study highlights the effectiveness of machine learning in identifying Alzheimer's risk post-COVID-19.

Conclusions:

• The Random Forest algorithm is the most effective model for predicting Alzheimer's disease in COVID-19 patients among the evaluated algorithms.
• These findings offer a valuable tool for proactive management and potential prevention of Alzheimer's-related issues in individuals recovering from COVID-19.
• Early identification through predictive modeling can significantly improve patient outcomes and reduce the long-term burden of cognitive decline.