人工神经网络分析错畸形减数矫治相关因素的研究

目的分析错畸形矫治减数与否的相关因素。方法建立用于是否减数预测的人工神经网络模型。结果采用23个指标:非量化指标2个,量化指标包括模型测量指标5个,头影测量硬组织指标11个,头影测量软组织指标5个。数据处理后得出:其中开唇露齿、IMPA(L1-MP)两个指标贡献大;而指标FMA(FH-MP)贡献小;中间几个指标贡献大小变化范围不是太大,并不代表严格的数量关系。结论用神经网络的方法分析正畸矫治减数与否的相关因素是可行的。

Using artificial neural network to analyze relative factors in extraction design of orthodontics of malocclusion

Objective To analyze the relative factors when designing whether a malocclusion patient needs extraction before orthodontic treatment or not . Methods Artificial intelligence method is employed to establish an artificial neural network model for forecasting if extraction is needed. Results 23 indexes were input among which 2 were non-quantitative indexes. The 21 quantitative indexes included 5 model analysis ones, 11 hard tissue cephalometric ones and 5 soft tissue cephalometric ones. The results after data processing showed that 2 indexes “anterior tooth protrusion with lip incompetence" and IMPA(L1-MP) mode made the biggest contributions, while FMA(FH-MP) mode very made a small contribution. The contributions of other several indexes varied a little. And the variation was not distinct. Conclusions It's feasible to employ the method of neural network to analyze the relative factors when designing whether a malocclusion patient needs extraction before orthodontic treatmentor or not.