Exploratory analysis using machine learning algorithms to predict pinch strength by anthropometric and...

Area of Science:

• Biomedical Engineering
• Machine Learning
• Human Factors Engineering

Background:

• Accurate prediction of hand function is crucial for ergonomic tool design and preventing musculoskeletal disorders.
• Machine learning (ML) offers potential for analyzing complex relationships between anthropometry, demographics, and hand function.

Purpose of the Study:

• To evaluate various ML algorithms for predicting hand function.
• To identify key socio-demographic and anthropometric predictors of hand pinch strength.

Main Methods:

• A cross-sectional study involving 7119 healthy Iranian participants (aged 10-89).
• Measurement of 17 hand-forearm anthropometric dimensions and three types of pinch strength.
• Application and evaluation of 12 ML classifiers using 21 features.

Main Results:

• Hand length, stature, age, thumb length, and index finger length were most predictive of pinch strength.
• K-nearest neighbor, AdaBoost, and random forest classifiers achieved high accuracies (up to 96.75%) in predicting pinch types.
• Significant correlation found between anthropometric dimensions, socio-demographic factors, and hand function.

Conclusions:

• ML models can effectively predict hand function based on anthropometric and socio-demographic data.
• Identifying predictive factors can inform the design of ergonomic tools to minimize hand strain.
• This approach may help reduce the incidence of hand-related musculoskeletal disorders.