论文部分内容阅读
本文通过对欧佩克石油价格建立加法分解模型:Yt=Tt+St+Ct+Rt,其中t表示周数,(Tt)是趋势项,(St)是周期项,(Ct)是周期项,(Rt)是随机干扰项。分别对趋势项、季节项、周期项采用幂指数回归、傅立叶谱分解和自回归移动平均方法进行建模,对随机项进行正态性、白噪声等检验。用建立的模型对欧佩克周收盘价格历史数据进行拟合,拟合的相关指数R2=0.9247;对2007年7、8、9月的周收盘价进行动态预测,预测效果较好,所收集的7月、8月及9月初的实际周收盘价均落在95%置信区间内。
In this paper, an additive decomposition model is established for OPEC oil price: Yt = Tt + St + Ct + Rt, where t is the number of weeks, (Tt) is the trend term, (St) is the periodic term, ) Is a random disturbance term. The trend items, seasonal items, and periodic items are respectively exponentially-exponentially regression, Fourier spectral decomposition and autoregressive moving average methods. The random items are tested for normality and white noise. Fitting the historical data of OPEC weekly closing price with the established model, the fitting correlation index R2 = 0.9247; forecasting the weekly closing prices of July, August and September in 2007, the forecasting effect is better. The collected 7 The actual closing prices for the month, August and early September all fall within the 95% confidence interval.