Prognosis Prediction of Sudden Sensorineural Hearing Loss Using Ensemble Artificial Intelligence Learning Models

Area of Science:

• Otolaryngology
• Medical Informatics
• Machine Learning

Background:

• Sudden sensorineural hearing loss (SSNHL) prognosis is complex.
• Predictive models can aid in treatment decisions and patient management.
• Ensemble learning offers potential for improved predictive accuracy.

Purpose of the Study:

• To develop and evaluate ensemble learning models for SSNHL prognosis.
• To identify key predictive variables for SSNHL outcomes.
• To assess the AdaBoost algorithm's performance against other models.

Main Methods:

• A retrospective study of 1,572 SSNHL patients.
• Utilized four key variables: age, days after onset, vertigo, and hearing loss type.
• Compared AdaBoost, boosting, bagging, and stacking algorithms.

Main Results:

• AdaBoost achieved the highest accuracy (82.89%), precision (86.66%), F1 score (89.20), and AUC (0.79).
• Younger age, fewer days after onset, and less vertigo correlated with better outcomes.
• Age and days after onset were identified as critical predictive parameters.

Conclusions:

• The AdaBoost model demonstrates high efficacy in predicting SSNHL.
• Simple variables enhance the model's practicality for remote healthcare.
• Age is a significant factor influencing SSNHL prognosis.