RBF神经网络改进模型在医学生UPI中的仿真研究

Simulating Study of RBF Neural Network Improved Model in UPI of the Medical Students

采用概率竞争和RBF(Radial-based Function Method)相结合的神经网络模型,研究UPI(UniversityPersonality Inventory大学生心理健康清单)多变量非线性诊断指标和分类标准的内在联系。提出了一种UPI辅助诊断的概率竞争型RBF神经网络的解决方案,把UPI数据作为概率竞争和RBF神经网络改进模型的输入,构建2层改进的RBF神经网络仿真模型。实验结果显示,综合运用概率竞争和RBF神经网络的方法能使UPI分类达到较好的效果。构建的概率竞争神经网络模型用于UPI的辅助诊断是可行的。

With the neural network model combining probability competition and RBF（Radial-based Func- tion Method）, the research aims at studying the internal relation between the multi-variable nonqinear diag- nostic index and classification standard of UPI（University Personality Inventory）. A UPI aided diagnostic probability competition RBF neural network has been put forward. With the UPI as the input of probability competition and RBF neural network improved model, the two-layer modified RBF neural network simulation model has been constructed. The experimental result shows that with the combination of probability competi- tion and RBF neural network model, a better classification of UPI can be achieved. It is feasible to construct probability competition neural network model to be used as the UPI aided diagnosis.