摘要
本文基于 BP神经网络模型实现了 1999～ 2 0 0 5年数据通信业务的业务收入发展预测 ,对于预测中的不可量化影响因素 ,根据其发展变化的历史序列和预期状况进行了编码量化 ,实现了不可量化因素输入定量模型的量化处理。通过本问题的建模与预测结果表明 ,该方法具有很强的学习与泛化能力 ,对于处理相关回归预测中的不可量化因素具有很好的应用价值。
The letter is based upon the NN model and realize operation earning development forecast of data communication service form 1999 to 2005, to the influence complication that can not be quantized in forecasting. We code and quantize according to its annals list and anticipation status of its development changement, we realize quantification of complication that can not be quantified importing ration model. During modeling and forecasting outcomes of this question, the measure has very strong ability of study and widely, it has good applied cost to dealing complication that can not be quantized related regression forecasting.
出处
《预测》
CSSCI
2000年第6期60-64,共5页
Forecasting
关键词
数据通信业务收入
BP神经网络
不可量化因素
预测
data communication service income
BPNN
complication that can not be quantized
forecast
