基于粒子群改进BP神经网络的组合预测模型及其应用

来源期刊:中南大学学报(自然科学版)2009年第1期

论文作者:崔吉峰 乞建勋 杨尚东

文章页码:190 - 194

关键词:BP神经网络;粒子群算法;ARIMA模型;灰色理论;组合预测

Key words:BP neural network; particle swarm optimization; auto regression-moving average; grey model; combining forecasting

摘    要:针对应用广泛的传统人工智能预测BP(Back propagation)神经网络自身局限以及其在处理中长期复杂预测问题中需要样本数量大、泛化能力弱等不足,提出利用粒子群算法优化BP神经网络的学习算法,在此基础上,利用灰色预测方法和自回归移动平均模型(ARIMA)时序预测对历史数据进行初步预测,对中长期预测中数据趋势项和随机项进行模拟;将初步预测的结果作为改进BP神经网络的输入,在此基础上进行训练和预测,构建基于改进BP网络的组合预测模型。以我国1978—2007年能源需求数据为样本,进行实例分析。结果表明:组合预测模型预测精度较BP神经网络、灰色预测方法和ARIMA预测方法分别提高4.8%,6.1%和5.3%,验证了组合预测方法在中长期预测问题处理中的有效性。

Abstract: To solve the limitations of BP neural network in dealing with complex and long-term prediction, which needs many samples to calculate and its generalization ability is poor, the learning algorithm of BP neural network optimized by PSO (Particle swarm algorithm) was proposed. On the basis of this algorithm, the grey prediction and the ARIMA (Autoregressive integrated moving average model) time series forecasting methods were used to make a preliminary forecast for historical data. The results of initial forecasts were put as the input of improved BP neural network to be forecasted and trained. An improved BP network-based combining forecasting model was built. China’s energy demands data in 1978—2007 were taken as the sample to be analyzed. The results show that the learning algorithm of BP neural network optimized by PSO has better effects in simulation of the trend data and the random data in medium- and long- term forecasting. The prediction accuracy of combining forecasting model increases by 4.8%, 6.1% and 5.3% respectively compared with those of BP neural network, gray forecasting methods and ARIMA forecasting, which proves the effectiveness of combining forecasting method in the medium-term and long-term forecasting.

基金信息:国家自然科学基金资助项目

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号