دانلود رایگان مقاله لاتین مقیاس درصد خطای مطلق از سایت الزویر
عنوان فارسی مقاله:
مقیاس های جدید درصد خطای مطلق برای پیش بینی های تقاضای متناوب
عنوان انگلیسی مقاله:
A new metric of absolute percentage error for intermittent demand forecasts
سال انتشار : 2016
برای دانلود رایگان مقاله مقیاس درصد خطای مطلق اینجا کلیک نمایید.
بخشی از مقاله انگلیسی:
3. Properties of MAAPE
This section compares MAPE and MAAPE, in order to investigate the properties of MAAPE. Recall that APE and AAPE are defined by components of MAPE and MAAPE, as in Eqs. (1.1) and (2.2), respectively. Without loss of generality, we therefore compare APE and AAPE. Fig. 3 provides visualizations of APE and AAPE in the upper and lower rows, respectively, with actual (A) and forecast (F ) values that vary from 0.1 to 10 in increments of 0.1. In the left column, the values of each measure are presented in a color map, varying from blue (low values) to red (high values). The actual and forecast values are on the x- and y-axes, respectively. For example, in Fig. 3(a), the upper-left corner presents APE values for small actual values and large forecast values, while the lower-right corner presents APE values for large actual values and small forecast values. As expected, the APE values in the upperleft corner are much larger than those in other regions. In the right column, the values of each measure on the diagonal line of the corresponding figure in the left column (from upper-left to lower-right) are plotted. On the x-axis in Fig. 3(b), both actual (A) and forecast (F ) values are presented; for simplicity, the x-axis can be regarded as F /A. Fig. 3(a) and (b) clearly illustrate the drawbacks of MAPE: it provides extremely large values when the actual values are small. In contrast, it can be seen clearly in Fig. 3(c) and (d) that AAPE does not go to infinity even with close-to-zero actual values, which is a significant advantage of MAAPE over MAPE. It is evident from a comparison of Fig. 3(c) and (d) with Fig. 3(a) and (b) that AAPE is less sensitive to small actual values than APE.
برای دانلود رایگان مقاله مقیاس درصد خطای مطلق اینجا کلیک نمایید.
کلمات کلیدی:
another look at forecast-accuracy metrics for ... - Semantic Scholar https://pdfs.semanticscholar.org/6fdc/9d43ad105d3e68a319430cc3e7d60264d7df.pdf by RJ Hyndman - Cited by 107 - Related articles He also introduces a new metric—the mean absolute scaled error. (MASE)—which is more ... absolute error (MAE or MAD); percentage-error metrics such as the ... A new accuracy measure based on bounded relative error for time ... journals.plos.org/plosone/article?id=10.1371/journal.pone.0174202 by C Chen - 2017 Mar 24, 2017 - Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) can be considered as the very early and most popular accuracy ... Sparrho | A new metric of absolute percentage error for intermittent https://www.sparrho.com/item/a-new-metric-of-absolute-percentage-error-for.../9cf3cd/ Oct 14, 2016 - Publication date: July–September 2016 Source:International Journal of Forecasting, Volume 32, Issue 3 Author(s): Sungil Kim, Heeyoung Kim ... People also ask What is MAPE? What is accuracy in forecasting? What is the forecast error? What is bias in forecasting? Feedback Searches related to new metric of absolute percentage error symmetric mean absolute percentage error in excel mean relative absolute error mean absolute scaled error mean absolute percentage error excel how to calculate forecast error in excel forecast accuracy metrics mean percentage error what is a good mape
