The Fifth Column: The Case of the Unused Forecast or When Trying to Forecast What’s to Be, First Try Y = MX + B
暂无分享,去创建一个
Once upon a time, a young MS/OR analyst for an airline received an unusual request for assistance—he was to review the volume-rate-forecasting model for the airline's reservation system. This state-of-the-art model was used for analyzing and forecasting univariate-time-series and transfer-function data based on an autoregressive integrated moving average technique. The request was unusual because its two originators had, between them, a combined total of five advanced degrees in mathematics. Our hero was, naturally, a bit curious about why such a pair were requesting help from a person holding an OR degree from an institution of the quick and dirty persuasion. After they explained that they wanted him to “fine tune” their model by looking at the white noise, degrees of differencing, residual normality, and perturbation values, he was even more curious. His curiosity disappeared when they admitted, through clenched teeth, that they had been ordered to seek his help.