论文部分内容阅读
本文使用神经网络对变电站造价与其影响因素进行分析、预测。首先,运用模糊数学理论对变电站工程样本进行优选,选出与待估价工程类似的训练样本;然后,运用BP神经网络实现变电站工程造价和主要影响因素之间的复杂非线性映射,进而用已建的模型对工程进行造价预测;最后,通过比对实际值和预测值,验证所建模型的预测精度。
In this paper, the use of neural network on the substation cost and its influencing factors were analyzed and predicted. Firstly, the fuzzy mathematics theory is used to optimize the engineering samples of the substation, and the training samples which are similar to the project to be evaluated are selected. Secondly, the complex nonlinear mapping between substation project cost and main influencing factors is realized by BP neural network, Of the model of the project cost forecast; Finally, by comparing the actual value and the predicted value, verify the model prediction accuracy.