此實驗是在討論電磁波量測的精確度問題，所用到的天線有兩種:偶極天線(Dipole antenna)和寬頻天線(Broadband antenna)。一般來說，一場地在測量電磁波的時候，天線在不同頻率所記錄觀察到的測量值沒有超過理論值(ideal)加減4的範圍的話，這個場地較被認定為是一個合格標準的場地，否則，就是不合格(只要有一個值超出範圍就是不合格)。傳統上，都是使用偶極天線(Dipole antenna)來測量，有鑒於傳統的偶極(Dipole)天線在操作上的困難度和不精確度，此研究是想用統計的方法來探討是否可用寬頻(Broadband)天線取代傳統的偶極天線來作對場地標準與否的量測。在此研究中，將先介紹了實驗的程序與資料，然後配適迴歸模型，來預測在不同因子水準下，不同頻率的測量值(measurements)。之後根據資料分析所收集到的，我們考慮用轉折點模型(change point model)來修正所作之統計分析，以期找到最適切的統計模型，最後再比較在相同的因子水準之下，寬頻天線和傳統偶極天線測量的差異。

In this work, we discuss the accuracy of measurements for
electromagnetic. The two kinds of antenna we use are Dipole
antenna and Broadband antenna. In general, if the antenna
measurements we recorded at different frequencies do not exceed
the ideal value $pm 4$dB, we would regard this site as a
normalized site, otherwise it is not a normalized site(just a
measurement exceeds the range). Traditionally, all we use is
Dipole antenna, but due to difficulty of operation and inaccuracy
of Dipole antenna, we investigate by statistical methods if we may
use the Broadband antenna to replace the traditional Dipole
antenna to measure. First of all, we introduce the data and
procedure in the experiments, and fit a statistical regression
model to predict the measurements at different frequencies in
different test setups. Then, according to the data we collected,
use the change point models to modify the statistical models. Our
goal is to find a suitable statistical model for the
measurements. Finally, we compare the measurements of Broadband
antenna
with Dipole antenna in the other experimental conditions keep the same.

Abstract
List of Tables
List of Figures
1.Introduction...............................1
2.Fitting models.............................2
3.The change point models....................3
4.Comparison................................11
5.Conclusion................................14
References..................................17

[1]
Akira.S. (1990).
Fomulation of Normalized Site Attenuation in Terms of Antenna
Impedances.
{it IEEE Transactions on Electromagnetic Compatibility} {f 32}, no.4, 257-263.
[2]
Akira.S. (1992).
Correction Factors for Normalized Site Attenuation.
{it IEEE Transactions on Electromagnetic Compatibility} {f 34}, no.4, 461-470.

[3]
Cater, R.L. and Blight, B.J.N. (1981).
A bayesian change-point problem with an application to the
prediction and detection of ovulation in women.
{it Biometrics} {f 37}, 743-751.

[4]
Chernoff, H. and Zacks, S. (1964).
Estimation the current mean of a normal distribution which is
subjected to changes in time.
{it Annals Math. Statist} {f 35}, 999-1018.

[5]
Chung, B.L. (1995).
Estimating the number of change points in a sequence of
independent normal random variables.
{it Statistics and Probability Letters} {f 25}, 241-248.
[6]
Freeman, J.M. (1984).
Two-phase regression and goodness of fit.
{it Communication in Statistics, Theorey and Methods} {f
13}, 1321-1334.
[7]
Freeman, J.M. (1986).
An unknown change point and goodness of fit.
{it The Statistician} {f 35}, 335-344.
[8]
Hinkley, D.V. (1969).
Inference about the intersection in two-phase regression.
{it Biometrika} {f 56}, 495-504.
[9]
Hinkley, D.V. and Hinkley, E.A. (1970).
Inference about the change point in a sequence of random variables.
{it Biometrika} {f 57}, 1-17.
[10]
Hinkley, D.V. (1971).
Inference about the change point from cumulative sum tests.
{it Biometrika} {f 58}, 509-523.
[11]
Hinkley, D.V. (1972).
Time ordered classification.
{it Biometrika} {f 59}, 509-523.

[12]
Hudson, D.J. (1966).
Fitting segmented curves whose join points have to estimated.
{it J. Amer. Statist. Assoc.} {f 61}, 1097-1129.
[13]
Krishnaiah, P.R. and Miao, B.Q. (1988).
Review About Estimation of Change Points.
{it Handbook of Statistics} {f 7}, 374-402.
[14]
Levene, H.(1960).
Robust Tests for Equality of Variances.
{it Contributions to Probability and Statistics}, ed. I.
Olkin. Palo Alto, Calif.: Stanford University Press, 278-292.
[15]
Neter, J., Kutner, K.H., Nachtshem, C.J. and Wasserman, W. (1999).
{it Applied Linear Regression Models}, third edition,
McGraw-Hill Int, N.Y.
[16]
Page, E.S. (1954).
Continuous inspection schemes.
{it Biometrika} {f 41}, 100-114.
[17]
Page, E.S. (1955).
A test for a change in a parameter occurring at an unknown
point.
{it Biometrika} {f 42}, 523-527.
[18]
Page, E.S. (1957).
On problems in which a change in parameter occurs at an
unknown point.
{it Biometrika} {f 44}, 248-252.
[19]
Quandt, R.E. (1958).
The estimation of the parameter of a linear regression system
obeying two separate regimes.
{it J. Amer. Statist. Assoc.} {f 53}, 873-880.
[20]
Quandt, R.E. (1960).
Tests of the hypotheses that a linear regression system obeys
two separate regimes.
{it J. Amer. Statist. Assoc.} {f 55}, 324.
[21]
Quandt, R.E. and Ramsey, B. (1978).
Estimating mixtures of normal distributions and switching regression.
{it J. Amer. Statist. Assoc.} {f 73}, 730-738.
[22]
Robison, D.E. (1964).
Estimating for the points of intersection of two polynomial
regressions.
{it J. Amer. Statist. Assoc.} {f 59}, 214-224.
[23]
Ronald, D.S. (1977).
Validation of Regression Models: Methods and Examples.
{it Technometrics} {f 19}, 415-428.
[24]
Smith, A.F.M. (1975).
A Bayesian approach to inference about a change-point in a
sequence of random variables.
{it Biometrika } {f 62}, 407-416.