The multivariate linear errors-in-variables model when the regressors are missing at random in the sense of Rubin （1976） is considered in this paper. A constrained empirical likelihood confidence region for a parame...The multivariate linear errors-in-variables model when the regressors are missing at random in the sense of Rubin （1976） is considered in this paper. A constrained empirical likelihood confidence region for a parameter β0 in this model is proposed, which is constructed by combining the score function corresponding to the weighted squared orthogonal distance based on inverse probability with a constrained region of β0. It is shown that the empirical log-likelihood ratio at the true parameter converges to the standard chi-square distribution. Simulations show that the coverage rate of the proposed confidence region is closer to the nominal level and the length of confidence interval is narrower than those of the normal approximation of inverse probability weighted adjusted least square estimator in most cases. A real example is studied and the result supports the theory and simulation＇s conclusion.展开更多
Though EV model is theoretically more appropriate for applications in which measurement errors exist, people are still more inclined to use the ordinary regression models and the traditional LS method owing to the dif...Though EV model is theoretically more appropriate for applications in which measurement errors exist, people are still more inclined to use the ordinary regression models and the traditional LS method owing to the difficulties of statistical inference and computation. So it is meaningful to study the performance of LS estimate in EV model. In this article we obtain general conditions guaranteeing the asymptotic normality of the estimates of regression coefficients in the linear EV model. It is noticeable that the result is in some way different from the corresponding result in the ordinary regression model.展开更多
It is well known that when the random errors are iid. with finite variance, the week and the strong consiStency of LS estimate of multiple regression coefficients are equivalent. This note, by constructing a counter-e...It is well known that when the random errors are iid. with finite variance, the week and the strong consiStency of LS estimate of multiple regression coefficients are equivalent. This note, by constructing a counter-example, shows that this equivalence no longer holds true in case that the random errors possess only the r-th moment with 1≤5 T ＜ 2.展开更多
Let Yi=x'iβo+ei. 1≤i≤n, n≥1 be a linear regression model. Denote by βn the M-estimateof βo, using a convex function ρ. In [1], a basic theorem (Theorem A below) concerning the weakconsistency of βn. is est...Let Yi=x'iβo+ei. 1≤i≤n, n≥1 be a linear regression model. Denote by βn the M-estimateof βo, using a convex function ρ. In [1], a basic theorem (Theorem A below) concerning the weakconsistency of βn. is established. This theorem raises further questions concerning the consistencyof βn. In this note, some of these questions are considered for the special cases of LAD and LSestimates.展开更多
Recently, frequency-based least-squares （LS） estimators have found wide application in identifying aircraft flutter parameters. However, the frequency methods are often known to suffer from numerical difficulties wh...Recently, frequency-based least-squares （LS） estimators have found wide application in identifying aircraft flutter parameters. However, the frequency methods are often known to suffer from numerical difficulties when identifying a continuous-time model, especially, of broader frequency or higher order. In this article, a numerically robust LS estimator based on vector orthogonal polynomial is proposed to solve the numerical problem of multivariable systems and applied to the flutter testing. The key idea of this method is to represent the frequency response function （FRF） matrix by a right matrix fraction description （RMFD） model, and expand the numerator and denominator polynomial matrices on a vector orthogonal basis. As a result, a perfect numerical condition （numerical condition equals 1） can be obtained for linear LS estimator. Finally, this method is verified by flutter test of a wing model in a wind tunnel and real flight flutter test of an aircraft. The results are compared to those with notably LMS PolyMAX, which is not troubled by the numerical problem as it is established in z domain （e.g. derived from a discrete-time model）. The verification has evidenced that this method, apart from overcoming the numerical problem, yields the results comparable to those acquired with LMS PolyMAX, or even considerably better at some frequency bands.展开更多
基金supported by the Natural Science Foundation of China under Grant Nos.10771017 and 11071022Key Project of MOE,PRC under Grant No.309007
文摘The multivariate linear errors-in-variables model when the regressors are missing at random in the sense of Rubin （1976） is considered in this paper. A constrained empirical likelihood confidence region for a parameter β0 in this model is proposed, which is constructed by combining the score function corresponding to the weighted squared orthogonal distance based on inverse probability with a constrained region of β0. It is shown that the empirical log-likelihood ratio at the true parameter converges to the standard chi-square distribution. Simulations show that the coverage rate of the proposed confidence region is closer to the nominal level and the length of confidence interval is narrower than those of the normal approximation of inverse probability weighted adjusted least square estimator in most cases. A real example is studied and the result supports the theory and simulation＇s conclusion.
文摘Though EV model is theoretically more appropriate for applications in which measurement errors exist, people are still more inclined to use the ordinary regression models and the traditional LS method owing to the difficulties of statistical inference and computation. So it is meaningful to study the performance of LS estimate in EV model. In this article we obtain general conditions guaranteeing the asymptotic normality of the estimates of regression coefficients in the linear EV model. It is noticeable that the result is in some way different from the corresponding result in the ordinary regression model.
基金the National Natural Science Foundation of China (No.19631040).
文摘It is well known that when the random errors are iid. with finite variance, the week and the strong consiStency of LS estimate of multiple regression coefficients are equivalent. This note, by constructing a counter-example, shows that this equivalence no longer holds true in case that the random errors possess only the r-th moment with 1≤5 T ＜ 2.
文摘Let Yi=x'iβo+ei. 1≤i≤n, n≥1 be a linear regression model. Denote by βn the M-estimateof βo, using a convex function ρ. In [1], a basic theorem (Theorem A below) concerning the weakconsistency of βn. is established. This theorem raises further questions concerning the consistencyof βn. In this note, some of these questions are considered for the special cases of LAD and LSestimates.
基金Foundation items： Aeronautical Science Foundation of China （2007ZD53053） NPU Foundation for Fundamental Research （NPU-FFR-W018104）
文摘Recently, frequency-based least-squares （LS） estimators have found wide application in identifying aircraft flutter parameters. However, the frequency methods are often known to suffer from numerical difficulties when identifying a continuous-time model, especially, of broader frequency or higher order. In this article, a numerically robust LS estimator based on vector orthogonal polynomial is proposed to solve the numerical problem of multivariable systems and applied to the flutter testing. The key idea of this method is to represent the frequency response function （FRF） matrix by a right matrix fraction description （RMFD） model, and expand the numerator and denominator polynomial matrices on a vector orthogonal basis. As a result, a perfect numerical condition （numerical condition equals 1） can be obtained for linear LS estimator. Finally, this method is verified by flutter test of a wing model in a wind tunnel and real flight flutter test of an aircraft. The results are compared to those with notably LMS PolyMAX, which is not troubled by the numerical problem as it is established in z domain （e.g. derived from a discrete-time model）. The verification has evidenced that this method, apart from overcoming the numerical problem, yields the results comparable to those acquired with LMS PolyMAX, or even considerably better at some frequency bands.
