本帖最后由 lanj 于 2009-12-24 09:12 编辑
Analyzing the results分析结果 CB Predictor simplifies the actual forecasting process, but you still must understand the results it gives you. CB Predictor对预测的过程进行了简化,但是你仍然需要了解其得到的结果。 Pasted forecasts粘贴结果 Two settings in the Results tab affect pasted data: the Paste Forecasts setting and the number of periods to forecast. If you selected Paste Forecasts, when you click Run, CB Predictor pastes the number of forecasted values at the end of your historical data or somewhere else if you specify another location. The worksheet automatically scrolls to where the pasted data begin. The forecasted values can appear in bold or italics, depending on the selected settings in the Paste Preferences dialog. 粘贴数据可以在结果表的两个地方进行设置:粘贴预测设置和预测周期数。如果在运行结果的时候选择的是粘贴预测结果,CB Predictor将在历史数据的后面或者你选择的其他的位置粘贴预测数据。工作表会自动定位到粘贴数据的开始位置。 预测值可以显示为粗体或者斜体,这可以在粘贴图表属性中进行设置。 Even through CB Predictor tried all the methods you selected in the Method Gallery, it generates the pasted values using the best method, unless you overrode the best method. In that case, the program generates the pasted values using the overriding method. CB Predictor可以根据你在方法图表中的选择尝试所有的预测方法,即使你没有选择最佳的方法,它也会用最佳的方式粘贴数值。一般默认的情况是使用覆盖的方式进行粘贴。 Statistical Note: Of the eight time-series forecasting methods, two result in flat lines: single moving average and single exponential smoothing. The forecasted values for these are all the same. This isn’t an error. It is the best possible forecast for volatile or patternless data. 注意:在八种预测方法中,有两种的结果是直线:单边移动平均和单边指数平滑。这两种的方法得到的结果是一样的,这并不是运算错误,而且因为这是波动数据和平稳数据最可能出现的情况。 When you paste regression results, the independent variable forecast values are pasted as simple value cells. The dependent variable forecasted values are created as formula cells with the regression equation as the formula. The regression equation coefficients appear below the pasted values. 当你粘贴回归结果的时候,粘贴的因变量的预测值只是个简单的数值,自变量的预测值是回归方程中得到的方程公式。回归方程的系数显示在粘贴值的下面。 If you did not generate any charts with your results, you can create an Excel chart by selecting the forecasted values ( along with some historical data) and using Excel’s Chart wizard. 如果没有生成任何的图表结果,你可以通过选择预测值和excel图表工具来制作excel图表。 Charts图表 Three settings in the Results tab affect charts:
在图表中三种设定结果:
Number Of Periods To Forecast 预测的周期数 | Sets how many forecasted values appear on the chart. 设定在报告中显示多少预测点 | Confidence Interval 置信区间 | Sets which confidence interval to calculate. If you select None, you cannot display a confidence interval on your charts. 设定预测结果的计算置信区间,如果你没有选择,在预测结果图表中设定置信区间 | Chart 图表 | Selects whether to generate charts. 选择是否给出图表 |
If you select Chart and click Run, CB Predictor generates Excel charts for each series with the number of forecasted values. The charts appear either on a different worksheet of the current workbook or on a different workbook, depending on the selected settings in the Chart Preferences dialog. The data used to generate the charts appears to the right of the charts. 如果选择了得到图表,CB Predictor会生成关于每列预测数据的excel图表。图表会在另外的sheet表或者新的excel工作表中生成,这可以在图表属性对话框中进行选择。数据一般会在图表的右边生成。 Even though CB Predictor tries all the methods you selected in the Method Gallery, it generates the chart values using the best method, unless you overrode the best method. In that case, the program generates the chart values using the overriding method. CB Predictor可以根据你在方法图表中的选择尝试所有的预测方法,但是一般情况它会选择最佳的方法,除非你不选择这个方法。一般情况,系统默认的是使用覆盖的方法得到预测值。 On the chart, the green line represents your historical data, the blue lines represent both the fitted and forecasted values, and the red lines above and below the forecasted values represent the upper and lower confidence interval. There is a gap between where the historical and fitted values end and the forecasted values begin to delineate the past and future values. 在图表中,绿色的线条代表历史数据,蓝色的线条代表最佳值和预测值,预测值上下的红色线条代表置信区间的上下限。在历史数据和最佳值的最后和预测值之间有间隔,以此来区分历史和预测的数值。 Statistical Note: Of the eight time-series forecasting methods, four result in straight lines: single moving average, single exponential smoothing, double moving average, and Holt’s double exponential smoothing. Only the seasonal methods and multiple linear regression result in curves that try to approximate any repeated data pattern. 注意:在八种预测方法中,有几种的结果是直线:单边移动平均和单边指数平滑,双边移动平均和Holt’s双边指数平滑。只有周期性方法和多元线性回归到结果是曲线形式。
-----------------------------------------------------------------------
译者权归“思步网(www.step365.com)”及“lanj(西北偏北)”共同所有,未经许可,请勿转载 |