Abstract

This research studies factors and creates a model to predict the patients’ postoperative WOMAC score after total knee replacement. First, the influencing factors were found by feature engineering, using several techniques such as Generalized Linear Models, Support Vector Machines, Deep Learning, and Gradient Boost Trees. Afterwards, the model was created by the Gradient Boost Tree technique which groups different attributes from feature engineering. Models were compared to find the model with the best predictability. RapidMiner Studio software version 9.9 was used in this work. The results demonstrate that the model created by the Gradient Boost Tree technique with attributes originating from feature engineering on the Gradient Boost Tree performs most efficiently with root mean square error (RMSE), mean absolute deviation (MAD) and square error (SE) of \mathbf{5}.\mathbf{311}\pm\mathbf{0}.\mathbf{538}, \mathbf{3}.\mathbf{550}\pm\mathbf{0}.\mathbf{376}, and \mathbf{28}.\mathbf{472}\pm\mathbf{5}.\mathbf{811} respectively.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call