Optimization enabled ensemble based deep learning model for elderly falling risk prediction.

Area of Science:

• Gerontology and Biomedical Engineering
• Artificial Intelligence in Healthcare

Background:

• Fall risk in the elderly is a major concern, impacting safety and well-being.
• Aging and chronic conditions frequently compromise balance, elevating the likelihood of falls.

Purpose of the Study:

• To develop an advanced fall risk prediction model utilizing an optimized deep learning methodology.
• To enhance the accuracy and reliability of fall risk assessment in the elderly population.

Main Methods:

• Data pre-processing and augmentation were employed to expand the dataset.
• An ensemble learning model integrating Extreme Gradient Boosting (XGBoost), One Dimensional Convolutional Neural Network, and Deep Belief Network was utilized.
• A novel Double Exponential Lyrebird Optimization (DELOA) algorithm was developed and applied for model training.

Main Results:

• The DELOA-based ensemble learning model demonstrated superior performance compared to conventional methods.
• Experimental results confirmed the effectiveness of the proposed optimization algorithm in improving fall risk prediction.

Conclusions:

• The developed optimized deep learning model offers a promising tool for accurate fall risk prediction in the elderly.
• This approach has the potential to significantly contribute to fall prevention strategies and improve geriatric care.