CLINICAL DECISION FOR STRABOTOMY BASED ON IMPROVED NONLINEAR MIXTURE OF EXPERTS NEURAL NETWORKS
An improved nonlinear mixture of experts model (ME)provides a modular approach wherein component neural networks are made specialists on subparts of a problem. This paper studied the application of improved ME variants to multivariate nonlinear systems of clinical decision problems,which are known to be difficult to be dealt with. The aim is to develop a new operation quantity planning decision model(OQPDM) based on improved nonlinear mixture of expert neural networks to predict the corrective quantity of lateral and medial rectus in strabotomy to instruct and improve practice. The corrective rate of strabotomy from OQPDM(97％) is better than past experience (76％) and shows effective prediction. OQPDM with improved nonlinear ME can offer robustness for potential application in other clinical decision support system, which can be implemented to develop a embeddable system for minisized instrument for eye checking.
Experts System clinical decision neural networks
WEI WANG LAN-FENG YAN BAO-WEI LIU YAN-JUN SHI
Institute of Biomedical Engineering, Key Lab of New Drugs Preclinical Study of Gansu, Lanzhou Univer Peoples Hospital of Gansu Province, West Dong-gang Road, 730000,China
国际会议
2006 International Conference on Machine Learning and Cybernetics(IEEE第五届机器学习与控制论坛)
大连
英文
2329-2334
2006-08-13（万方平台首次上网日期，不代表论文的发表时间）