摘要
采用传统Delphi方法进行决策、预测时,成员权重主要根据成员的经历、职务、年龄和自我评定等情况来确定,易导致加权平均值及方差计算不准确,影响Delphi法的精度和效率。为提高传统Delphi法的精度和效率,本文在基于BP神经网络的加权平均值计算模型基础上,提出了基于BP神经网络的改进Delphi法。改进的Delphi法使成员权值分配与其决策预测结果直接相关,减少了人为不正确因素对权值分配的影响。该方法已被应用于股票上市公司经营业绩综合评价排序,通过与传统Delphi法应用对比,证实了改进的Delphi法具有较高的计算精度和效率。
As member＇ weight in the traditional Delphi method is mainly determined by his experiences, position, age and evaluation etc, it is possible to lower the accuracy for making decision and forecast. To enhance the accuracy of the traditional Delphi method, an improved Delphi method based on BP artificial neural network is proposed. By establishing a new weight calculation model based on BP artificial neural network for computing member＇ weight, the improved Delphi method can make decision and forecast accurately. It has been successfully applied to evaluate and sort operational results of stock listed companies of China, the result shows that the improved Delphi method is more accurate than the traditional Delphi method.
出处
《微计算机信息》
北大核心
2006年第10X期171-173,共3页
Control & Automation
基金
国家自然科学基金资助项目(编号:70361002)
