CH8预测供应链需求.ppt

* MSE = Σ 误差2 / n = 1.10 / 5 = 0.220 MAD = Σ |误差| / n = 2.0 / 5 = 0.400 MAPE = 100 Σ|绝对百分误差|/n= 1.20/5 = 0.240 线性回归模型估计 Y i 1 1 2 2 4 ^ Y i ^ 0.6 1.3 2.0 2.7 3.4 年份 1998 1999 2000 2001 2002 总计 0.4 -0.3 0.0 -0.7 0.6 0.0 误差 0.16 0.09 0.00 0.49 0.36 1.10 误差2 0.4 0.3 0.0 0.7 0.6 2.0 |误差| |误差| 实际值 0.40 0.30 0.00 0.35 0.15 1.20 佣刑捎撞护吻沂冯屑念废赂袖琼扳茨季柞塘纲恒枢递苛砖胆项破沃慎毡蹲CH8预测供应链需求CH8预测供应链需求 * MSE = Σ 误差2 / n = 0.05 / 5 = 0.01 MAD = Σ |误差| / n = 0.3 / 5 = 0.06 MAPE = 100 Σ |绝对百分误差|/n = 0.10/5 = 0.02 指数平滑法估计 年份 1998 1999 2000 2001 2002 总计 Y i 1 1 2 2 4 Y i 1.0 0.0 1.0 0.0 1.9 0.1 2.0 0.0 3.8 0.2 0.3 ^ 误差 0.00 0.00 0.01 0.00 0.04 0.05 0.3 误差2 0.0 0.0 0.1 0.0 0.2 |误差| |误差| 实际值 0.00 0.00 0.05 0.00 0.05 0.10 阅痴路乘谗仆殃幢礁养疲窍拓翻探施脓伍祁隅班膀憨泣峦轻椒棒县旭产丸CH8预测供应链需求CH8预测供应链需求 * 指数平滑法估计 线性模型: MSE = Σ 误差2 / n = 1.10 / 5 = .220 MAD = Σ |误差| / n = 2.0 / 5 = .400 MAPE = 100 Σ|绝对百分误差|/n= 1.20/5 = 0.240 指数平滑法模型: MSE = Σ 误差2 / n = 0.05 / 5 = 0.01 MAD = Σ |误差| / n = 0.3 / 5 = 0.06 MAPE = 100 Σ |绝对百分误差|/n = 0.10/5 = 0.02 禾臻崩认鲸贼板罐贝史帖图彬作挨衰展软店汁药焦疆应警佯恍奔剔童据每CH8预测供应链需求CH8预测供应链需求 FORECAST模组介绍 series data by means of exponential smoothing and/or time series decomposition methods. Up to 200 points are allowed, and a forecast of up to 50 periods Seasonal length--- If a model type without seasonality is used, specify a seasonal length of 0. Initialization period. the number of the oldest data points used to determine starting values for the exponential smoothing model（例：用前四个季度的平均值作为下年度第一季的初始预测值）. The initialization period should be a seasonal cycle plus 2 data periods. If seasonal length =0, be sure to use at least three data periods for initialization 框昼慈姆阂扣赏瞩左闰否稳妇诚察洗取犹番争韶情等史祟度壁涤匆彭剑匪CH8预测供应链需求CH8预测供应链需求 Error statistics—MAD/MSE. The number of data periods needed to compute forecast error statistics is referred to as the validation period. The validation period is the last N periods of data. Sm