Affiliations: College of Mechanical and Electrical Engineering, Harbin Engineering University, Harbin, Heilongjiang, China
Correspondence:
[*]
Corresponding author: Yan Xu, College of Mechanical and Electrical Engineering, Harbin Engineering University, Harbin 150001, Heilongjiang, China. Tel.: +86 18746052730; E-mail:wangrilian09@163.com
Abstract: BACKGROUND: Recently, acupuncture of Traditional Chinese
Medicine (TCM) becomes a better method compared with the western medicine
for treating eye movement disorder. However, because therapeutic mechanism
of TCM is very complicated, acupuncture of TCM depends on the doctor's
experience to a large extent. INTRODUCTION: A Decision Support System (DSS) is urgently needed to
carry out Computer Aided Diagnosis (CAD) in order to provide doctors with
required clinical reference. MATERIALS AND METHODS: Save the data obtained by diplopia image
test software to the database, and then upload the data from various data
centers to the DSS which is based on Support Vector Machine (SVM) optimized
by Particle Swarm Optimization (PSO) through the Web interface. Train these
data to obtain a classification model with high accuracy, then download this
classification model back to every data center to help them make
classification prediction for new test samples, realizing CAD of eye
movement disorder. RESULTS: The test software can get the diplopia images and test
data quickly, and then the test data can be successfully stored to the
database for future retrieve. Moreover the DSS can achieve the CAD with good
results. DISCUSSION: The rounded software system can realize its expectant
functions. It can get accurate classification model quickly. The accuracy of
classification is increased to 62.5%. CONCLUSIONS: The feasibility and accuracy of this DSS was proved by
clinical experiments with good results. So it can be applied to provide CAD
for eye movement disorder in hospitals.
Keywords: Web based decision support system, eye movement disorder, Traditional Chinese Medicine (TCM), Support Vector Machine (SVM), Particle Swarm Optimization (PSO)
