同时针对神经网络易于陷入局部极值、结构难以确定和泛化能力较差的缺点，引入了能很好解决小样本、非线性和高维数问题的支持向量回归机来进行油气田开发指标的预测；
The method of support vector regression which can well resolve the problem with the insufficient swatch, nonlinear and high dimension is introduction to predict the development index of gas-field.
针对信息科学和控制理论中经常涉及的一类泛函极值问题，提出基于连续回归神经网络的求解方法。
In this paper, the continuous time recurrent neural network is proposed to solve the functional minimization problem, which is often involved in estimation and control.
与多元线性回归、模糊回归和自适应模糊神经网络相比，该模型学习精度高且具有较好的泛化能力，能取得较好的预测效果。
Comparing with the models based on multiple statistic analysis, generalized regress-ion neural network or adapted fuzzy neural network model, it shows better learning precision and generalization.
应用推荐