根据样本空间的内积特性，提出一种无需迭代学习内积神经网络。以某型航空发动机的机载记录数据为例，对发动机进行了建模，结果表明该方法具有自学习速度快、抗干扰能力强、准确性高的特点。

According to the inner product feature of sample space,a new neural network which does not require to learn iteratively,is set up for a aeroengine in terms of data recorded on a plane.The results show that the new method has the advantage of faster self taught ability,higher accuracy,stronger anti interference ability and less maintenance work.

V233.75