當我們在測量一儀器或設備的電磁波強度時，一開放空間的EMI觀察值是最直接也是最常被接受的測量標準。天線的觀測值將決定一個場地是否適合用來測量 EMI。在本論文之中，我們將試著從不同因子組合下的資料來找出觀測值和天線狀態間的的關係。在此之前，一個轉折點的轉折點模型已被用來配適觀察值並比較兩種天線(寬頻天線和偶極天線)間的不同。然而，並非全部的資料都適合用一個轉折點的轉折點模型來配適。所以，我們亦用不同的模型來配適資料。此外， 我們利用統計推論的結果，找出比理論值 4dB還嚴格的標準來判斷場地之間的優劣。最後，我們用Matlab製作一個使用者介面來處理統計分析的工作，並試著使其成為一個能鑑別場地優劣的基本工具。

Open site measurement on the electromagnetic interference is the most direct and universally accepted standard approach for measuring radiated emissions from an equipment
or the radiation susceptibility of a component or equipment. A site is qualified for testing
EMI or not is decided by the antenna measurements. In this work, we use data from setups with di erent factors to find relations of measurement and the situation of antenna. A one change point model has been used to fit observed measurements and compare the di erences
with two kinds of antenna (broadband antenna and dipole antenna). However, with only one change point model it may not give a suitable fit for all data sets in this work.
Therefore, we have tried other models and applied them to the data. Furthermore, we try to set up another standard more strict than ±4dB based on statistical inference results in deciding whether a site is a better one with more precision in measuring EMI values.
Finally, a program by Matlab with a complete analysis based on the procedure performed here is provided, so that it may be used as a standard tool for evaluating whether a site is with good measurement quality in practice.

Abstract..............................ii
List of Tables .......................iii
List of Figures.......................iv
1 Introduction.......................1
2 Data description ..................2
3 Model Fitting and Prediction.......6
4 Results and Discussions............10
5 Conclusion.........................12
Reference.............................14
Appendix..............................14
1.The theoretical formula of EMAXD....14
2. The Wilcoxon Signed Ranks Test.....16
3. The Friedman Test..................17
4. Software...........................18

1. Akira, S. (1990). Formulation of Normalized Site Attenuation in Terms of Antenna
Impedances, IEEE Transactions on Electromagnetic Compatibility, 32 , 257-263.
2. Akira, S. (1992). Correction Factors for Normalized Site Attenuation, IEEE Transactions
on Electromagnetic Compatibility, 34, 461-470.
3. Page, E. S. (1954). Continuous inspection schemes, Biometrika, 41, 100-114.
4. Page, E. S. (1955). A test for a change in a parameter occurring at an unknown
point, Biometrika, 42, 523-527.
5. Page, E. S. (1957). On problems in which a change in parameter occurs at an
unknown point, Biometrika, 44, 248-252.
6. Hudson, D. J. (1966). Optimal multivariate designs, J. Amer. Statist. 61, 1097-1129.
7. Akaike, H. (1974). A new look at the statistical model identification, IEEE Transaction
on Automatic Control, AC-19, 716-723.
8. Hurvich, C. M. and Tsai, C. -L. (1989). Regression and time series model selection
in small samples, Biometrika, 76, 297-307.
9. Gideon S. (1978) Estimating the dimension of a model, The Annals of Statistics, 5,
461-464.
10. Richard H. J. and Indranil D. (1995). Determining one or more change points,
Chemistry and Physics of LIPIDS, 76, 1-6.
11. Wang, D.Y. , Lin, K.-H. , Huang, M.-N. L. , and Cheng, C.P. (2004). Uncertainty
Studies of Radiated Emissions Test Site Based on Measured Normalized Site Attenuation
Data. manuscript.