Healthy Knee Kinematic Phenotypes Identification Based on a Clustering Data Analysis

The purpose of this study is to identify healthy phenotypes in knee kinematics based on clustering data analysis. Our analysis uses the 3D curves, namely, flexion/extension, abduction/adduction, and tibial internal/external rotation, measured via a KneeKG™ system during gait task. We investigated two representation approaches that are joint three dimensions. first global approach considered concatenation kinematic without any dimensionality reduction. second local set 69 biomechanical parameters interest extracted from curves. representations followed by process, BIRCH (balanced iterative reducing using hierarchies) discriminant model, separate into homogeneous groups or clusters. Phenotypes were obtained averaging those groups. validated clusters inter-cluster correlation statistical hypothesis tests. simulation results showed more efficient, it allows identification descriptive within population.