This paper proposes such a methodology. It involves nonpara-metric statistical analyses as well as the determination of lower bounds to serve as bases for comparison. Three optimization methods¡ªthat of Hansmann and Rivett 1959 , a regression-based approach, and a neural network supported method¡ªare used to demonstrate the use of the methodology
The effectiveness of a bidding optimization method can be, and has been, gauged by comparing its profitability to that resulting from other methods. However, the manner in which these com-parisons have been conducted can result in some misleading con-clusions, typically overstating the effectiveness of a proposed method. Furthermore, in such comparisons no attempt is made to identify any theoretical upper bound on profitability.

本文提出这样方法学。 它介入非参数统计分析并且最低界面的决心起基地作用对于比较。 三个优化方法Hansmann ¡ ªthat和Rivett 1959年，一种基于退化的方法和一个神经网络支持方法被用于的¡ ªare展示使用方法学 一个出价的优化方法的The有效率可以是和是，测量通过比较它的有利与起因于其他方法的那。 然而，这些比较被举办了的方式可能导致一些引入歧途的结论，典型地夸张一个提出的方法的有效率。 此外，在这样比较尝试没有被做辨认在有利的任何理论最高界面 。－－这是我在http://fanyi.cn.yahoo.com/translate_txt翻译的