兰州大学学报（自然科学版）
蘭州大學學報（自然科學版）
란주대학학보（자연과학판）
JOURNAL OF LANZHOU UNIVERSITY（NATURAL SCIENCES）
2004年
6期
55-58
,共4页
翟红林%陈晓峰%陈兴国%胡之德
翟紅林%陳曉峰%陳興國%鬍之德
적홍림%진효봉%진흥국%호지덕
径向基概率神经网络%小儿厌食症%偏最小二乘法
徑嚮基概率神經網絡%小兒厭食癥%偏最小二乘法
경향기개솔신경망락%소인염식증%편최소이승법
radial basis function probabilistic neural network%infancy anorexia%partial least squares
结合了径向基神经网络较强模式分类能力与概率神经网络运算简单的优点,提出了一种径向基概率神经网络模型,并应用于小儿厌食症的辅助诊断,通过对119例样本数据的处理,获得了92.4%的准确率.此外,偏最小二乘法的分析结果表明,Zn元素与小儿厌食症关系最为紧密.
結閤瞭徑嚮基神經網絡較彊模式分類能力與概率神經網絡運算簡單的優點,提齣瞭一種徑嚮基概率神經網絡模型,併應用于小兒厭食癥的輔助診斷,通過對119例樣本數據的處理,穫得瞭92.4%的準確率.此外,偏最小二乘法的分析結果錶明,Zn元素與小兒厭食癥關繫最為緊密.
결합료경향기신경망락교강모식분류능력여개솔신경망락운산간단적우점,제출료일충경향기개솔신경망락모형,병응용우소인염식증적보조진단,통과대119례양본수거적처리,획득료92.4%적준학솔.차외,편최소이승법적분석결과표명,Zn원소여소인염식증관계최위긴밀.
Based on a radial basis function probabilistic neural network model, which combined the powerful capability of the pattern classification of radial basis function neural network and the simple operation of probabilistic neural network, a new approach of assisted diagnosis for infancy anorexia was developed and applied to 119 samples, with an accuracy rate of 92%. In addition, the result of partial least squares analysis indicated that Zn was the most important element that was closely related to infancy anorexia..
