| Due to the faultiness of measurement resources, the influence of measurement conditions and the restriction of measurement staffs' competence, the value measured by instruments are disaccord with the true value, which means the measurement error are in existence. Hence, measurement error indwells throughout all of the scientific experiments and measurement activities. As early as the 18th century, the classical error theory came through sprout and development. Until the middle of 1870s, the analysis of measurement error and data processing gradually hold an important position in scientific experiment and the practice of production along with the formation and development of the modem error theories.This paper discusses the origin and sort of error theories and different representations of error. Then it works over some elementary thoughts and process of realization of measurement error model, where the author particularly analyses the enactment of model, estimate of parameters and true values. Afterward this article introduces a typical method--likelihood in the measurement error models. It detailedly expounds the influence of observational error on statistical inference, using different examples under the simulating methods to validate that measurement error would have an inaccurate impact to the discriminant analysis. Then this article puts forward the common methods and their mathematics models to eliminate the measurement error.Here, Author analyzes the robustness of the model and the essential of hypothesis test. Through defining a few conceptions, this text interprets theoretical evidence for discriminant influences.At last, the jackknife method is used to raise the accuracy of the measurement results. Through the instance, the conclusion is proved that jackknife method is better for reducing the measurement error. Although the calculation quantity compares greatly, this method will have further applications along with the improvement of calculator speed.
|
PDF Full Text Request