頁籤選單縮合
題名 | Performance Analysis of Linear Regression Systems Subject to Inefficiency= |
---|---|
作者 | Bellamine, Ilyes; Morita, Hiroshi; Ishii, Hiroaki; |
期刊 | Asia Pacific Management Review |
出版日期 | 20040600 |
卷期 | 9:3 民93.06 |
頁次 | 頁379-390 |
分類號 | 448.931 |
語文 | eng |
關鍵詞 | Non-Stochastic inefficiency function; Linear regression system; Maximum likelihood; Akaike information criterion; |
英文摘要 | We propose a procedure for analysis of the performance of linear systems subject to inefficiency that is defined as the lack of output due to a change of some supposed constant factors assigned firstly to the disturbance term. To estimate the inefficiency it is recommended to decompose the systematic term into two terms, one representing the effect of the important (controllable) factors and the other representing the effect of factors that have been considered as stable but effectively changed during the experiment. The latter term is called inefficiency and it is approximated by a multi-step function. The number of gaps (or steps) is identified through the Akaike information criterion, and the parameters of the inefficiency function and the linear model are obtained by maximizing the appropriate likelihood function. The procedure enables us to decompose the levels of input into zones where the performance of the system is different. The paper concludes with a worked case study along with the performance of a mail processing process is analyzed. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。