常熟理工学院学报
常熟理工學院學報
상숙리공학원학보
JOURNAL OF CHANGSHU INSTITUTE OF TECHNOLOGY
2014年
2期
107-111
,共5页
工业废气排放量预测%交叉组合模型%灰色系统%前馈神经网络
工業廢氣排放量預測%交扠組閤模型%灰色繫統%前饋神經網絡
공업폐기배방량예측%교차조합모형%회색계통%전궤신경망락
industrial emissions forecast%combination model%gray system%back propagation neural network
城市工业废气排放量变化是非线性的,同时具有复杂的随机性和趋势性特点,传统单一预测模型难以对其变化规律进行准确表达,从而导致预测精度较低。为提高城市工业废气排放量的预测精度,提出了GM-BP组合模型。通过GM(1,1)模型对城市工业废气排放量变化趋势进行预测,然后运用BP神经网络模型对GM(1,1)模型的趋势预测值进行误差修正,以提高预测精度。对南京市2007~2010年城市工业废气排放量进行的仿真实验表明,GM-BP模型的预测精度较高,能够应用于城市工业废气排放量预测。
城市工業廢氣排放量變化是非線性的,同時具有複雜的隨機性和趨勢性特點,傳統單一預測模型難以對其變化規律進行準確錶達,從而導緻預測精度較低。為提高城市工業廢氣排放量的預測精度,提齣瞭GM-BP組閤模型。通過GM(1,1)模型對城市工業廢氣排放量變化趨勢進行預測,然後運用BP神經網絡模型對GM(1,1)模型的趨勢預測值進行誤差脩正,以提高預測精度。對南京市2007~2010年城市工業廢氣排放量進行的倣真實驗錶明,GM-BP模型的預測精度較高,能夠應用于城市工業廢氣排放量預測。
성시공업폐기배방량변화시비선성적,동시구유복잡적수궤성화추세성특점,전통단일예측모형난이대기변화규률진행준학표체,종이도치예측정도교저。위제고성시공업폐기배방량적예측정도,제출료GM-BP조합모형。통과GM(1,1)모형대성시공업폐기배방량변화추세진행예측,연후운용BP신경망락모형대GM(1,1)모형적추세예측치진행오차수정,이제고예측정도。대남경시2007~2010년성시공업폐기배방량진행적방진실험표명,GM-BP모형적예측정도교고,능구응용우성시공업폐기배방량예측。
The urban industrial emission which was influenced by many factors and its variation presented a se-ries of characters: nonlinear, trend and volatility, so that a single model cannot describe its changing rule accu-rately. GM-BP model combination of GM (1, 1) and BP neural network was proposed. GM (1, 1) model was used to simulate the sample data, and then BP neural network was trained on the prediction errors of GM (1, 1) and related factors of urban industrial emission. Finally, the sum of trend predicted value of GM (1, 1) and fluctua-tions prediction value of BP neural network were regarded as the prediction value of the GM-BP model. The ex-perimental results indicated that GM-BP model can increase the prediction accuracy and describe the changing rule of the urban industrial emission.