Orthopedic disease classification based on breadth-first search algorithm

Area of Science:

• Orthopedics
• Medical Diagnostics
• Machine Learning

Background:

• Orthopedic diseases are prevalent globally, affecting the musculoskeletal system and leading to pain and functional impairment.
• Current diagnostic methods for orthopedic conditions may lack sufficient supplementary tools.
• There is a need for improved and accurate methods for diagnosing orthopedic diseases.

Purpose of the Study:

• To address the limitations in supplementary diagnostics for orthopedic diseases.
• To enhance the methodology for diagnosing orthopedic diseases using advanced computational techniques.
• To evaluate the efficacy of various feature selection and machine learning models in orthopedic diagnosis.

Main Methods:

• Feature selection was performed using binary breadth-first search (BBFS), binary particle swarm optimization (BPSO), binary grey wolf optimizer (BGWO), and binary whale optimization algorithm (BWAO).
• Six machine learning models, including Random Forest (RF), Support Grained Descent (SGD), Naive Bayes Classifier (NBC), Discriminant Analysis (DC), Quadratic Discriminant Analysis (QDA), and Extra Trees (ET), were applied.
• The RF model's parameters were further optimized using BBFS, BPSO, BGWO, and BWAO, with performance evaluated using accuracy, sensitivity, specificity, F-score, and AUC curve.

Main Results:

• BBFS demonstrated an average error reduction of 47.29% compared to other feature selection algorithms.
• The Random Forest (RF) model initially achieved an accuracy of 91.4% among the tested machine learning models.
• The optimized BBFS-RF model significantly improved performance, reaching an accuracy of 99.41% on the dataset.

Conclusions:

• The integration of BBFS for feature selection and RF for classification offers a highly accurate approach to diagnosing orthopedic diseases.
• Optimized machine learning models, particularly BBFS-RF, show substantial promise in enhancing the accuracy and reliability of orthopedic diagnostics.
• This study provides a robust computational framework that can potentially improve patient outcomes through earlier and more precise disease detection.