Knee Joint Pathology Screening Using Time-Domain Multidimensional Fusion Feature and Random Forest

Knee-joint VAG signal analysis has a significant role in achieving early pathological screening of the knee joint and can be an efficient method for performing a non-invasive knee osteoarthritis (KOA) diagnosis. To improve the diagnostic accuracy of KOA, we presented a KOA pathology screening method based on time-domain multidimensional fusion feature (TDMFF) with the random forest, using feature fusion method to obtain a time-domain multidimensional fusion feature model to describe the fusion feature of VAG signals, combined with the random forest machine learning classifier for pathology screening. The research in this paper was verified by experiment results with collection testee’s normal and abnormal VAG signals. The KOA screening results illustrate that our classification has accuracy of 0.93, sensitivity of 0.93, precision of 0.93, and Fi-score of 0.93. The research results have a high screening rate for knee joint pathological screening and offer a novel practical way for non-invasive KOA screening.