A hybrid intelligent system combining self organizing maps and case based reasoning for evaluating postural control

Evaluating performance is crucial in stability assessments and training as it provides a valuable measure to examine postural control and recommend rehabilitation strategies for improvement. This paper proposes a novel methodology combining self organizing maps and case based reasoning implemented using a relational database management system to access, examine, diagnose and propose recommendations for improving trunk postural control. A self organizing map was developed to classify the input dataset gathered during a tandem Romberg stability test which behaved as the knowledgebase of the system. Case based reasoning was incorporated based on its ability to solve new problems by adopting previously defined, successful solutions to analogous problems. The case based reasoning architecture was implemented using a relational database with the self organizing map integrated to retrieve analogous case records to solve a new case. The prediction accuracy of the hybrid system to accurately produce the most similar case for a given new case was tested using leave-one-out-cross-validation method. The results demonstrated a high prediction accuracy of over 90% confirming the effectiveness of the hybrid methodology for evaluating and predicting recommendations for performance improvement based on existing knowledgebase for postural control.