Categories
Change Password!
Reset Password!
PSO-optimized neural networks achieve up to 100% accuracy in predicting musculoskeletal pain risk, highlighting core determinants like age, BMI, exercise, and occupational factors for targeted prevention.
Musculoskeletal pain is a prevalent health challenge worldwide. It not only influences physical well-being but also diminishes productivity and quality of life. Scientists are now turning to artificial intelligence (AI) and optimization techniques to better comprehend pain risk and design preventive strategies. This study set out to create and test a novel framework that unites particle swarm optimization (PSO) with neural networks to estimate musculoskeletal pain risk.
By analyzing demographic, occupational, physical, and lifestyle data, the goal was to uncover hidden risk patterns and identify key predictors of pain. Researchers utilized a dataset of 350 participants, which contained comprehensive information on pain experiences across multiple body regions. The data underwent preprocessing to address missing values, normalize features, and balance class distributions using the synthetic minority over-sampling technique (SMOTE). A feedforward neural network with one hidden layer was then designed, with PSO applied to optimize the network’s weights and biases.
The model’s accuracy and robustness were tested by standard metrics like precision, recall, F1-score, and AUC-ROC. The PSO-powered neural network achieved outstanding performance, with accuracy levels ranging between 95.8% and 100%. Importantly, the model determined strong predictors of musculoskeletal pain risk, including age, body mass index (BMI), frequency of exercise, and occupational demands. Compared with conventional approaches, this advanced framework showed an excellent ability to detect risk patterns and yield actionable insights.
This research highlights the promise of optimization-driven AI models in musculoskeletal pain assessment. By accurately predicting pain risk and pinpointing contributing factors, the framework paves the way for personalized interventions and preventive health strategies. Such approaches could help mitigate the global burden of musculoskeletal pain and boost overall quality of life.
Scientific Reports
Classification of musculoskeletal pain using machine learning
Dalia Mohamed Fouad et al.
Comments (0)