In this paper, we will provide a proposed method to estimate missing values for the Explanatory variables for Non-Parametric Multiple Regression Model and compare it with the Imputation Arithmetic mean Method, The basis of the idea of this method was based on how to employ the causal relationship between the variables in finding an efficient estimate of the missing value, we rely on the use of the Kernel estimate by Nadaraya – Watson Estimator , and on Least Squared Cross Validation (LSCV) to estimate the Bandwidth, and we use the simulation study to compare between the two methods.

In this study, we focused on the random coefficient estimation of the general regression and Swamy models of panel data. By using this type of data, the data give a better chance of obtaining a better method and better indicators. Entropy's methods have been used to estimate random coefficients for the general regression and Swamy of the panel data which were presented in two ways: the first represents the maximum dual Entropy and the second is general maximum Entropy in which a comparison between them have been done by using simulation to choose the optimal methods.

The results have been compared by using mean squares error and mean absolute percentage error to different cases in term of correlation valu

Abstract

This research aim to overcome the problem of dimensionality by using the methods of non-linear regression, which reduces the root of the average square error (RMSE), and is called the method of projection pursuit regression (PPR), which is one of the methods for reducing dimensions that work to overcome the problem of dimensionality (curse of dimensionality), The (PPR) method is a statistical technique that deals with finding the most important projections in multi-dimensional data , and With each finding projection , the data is reduced by linear compounds overall the projection. The process repeated to produce good projections until the best projections are obtained. The main idea of the PPR is to model

This research aims to study the methods of reduction of dimensions that overcome the problem curse of dimensionality when traditional methods fail to provide a good estimation of the parameters So this problem must be dealt with directly . Two methods were used to solve the problem of high dimensional data, The first method is the non-classical method Slice inverse regression ( SIR ) method and the proposed weight standard Sir (WSIR) method and principal components (PCA) which is the general method used in reducing dimensions,    (SIR ) and (PCA) is based on the work of linear combinations of a subset of the original explanatory variables, which may suffer from the problem of heterogeneity and the problem of linear

This research deals with unusual approach for analyzing the Simple Linear Regression via Linear Programming by Two - phase method, which is known in Operations Research: “O.R.”. The estimation here is found by solving optimization problem when adding artificial variables: Ri. Another method to analyze the Simple Linear Regression is introduced in this research, where the conditional Median of (y) was taken under consideration by minimizing the Sum of Absolute Residuals instead of finding the conditional Mean of (y) which depends on minimizing the Sum of Squared Residuals, that is called: “Median Regression”. Also, an Iterative Reweighted Least Squared based on the Absolute Residuals as weights is performed here as another method to

Abstract

      In this study, we compare between the autoregressive approximations (Yule-Walker equations, Least Squares , Least Squares ( forward- backword ) and Burg’s (Geometric and Harmonic ) methods, to determine the optimal approximation to the time series generated from the first - order moving Average non-invertible process, and fractionally - integrated noise process, with several values for d (d=0.15,0.25,0.35,0.45) for different sample sizes (small,median,large)for two processes . We depend on figure of merit function which proposed by author Shibata in 1980, to determine the theoretical optimal order according to min

The Dagum Regression Model, introduced to address limitations in traditional econometric models, provides enhanced flexibility for analyzing data characterized by heavy tails and asymmetry, which is common in income and wealth distributions. This paper develops and applies the Dagum model, demonstrating its advantages over other distributions such as the Log-Normal and Gamma distributions. The model's parameters are estimated using Maximum Likelihood Estimation (MLE) and the Method of Moments (MoM). A simulation study evaluates both methods' performance across various sample sizes, showing that MoM tends to offer more robust and precise estimates, particularly in small samples. These findings provide valuable insights into the ana

Mixture experiments are response variables based on the proportions of component for this mixture. In our research we will compare the scheffʼe model with the kronecker model for the mixture experiments, especially when the experimental area is restricted.

     Because of the experience of the mixture of high correlation problem and the problem of multicollinearity between the explanatory variables, which has an effect on the calculation of the Fisher information matrix of the regression model.

     to estimate the parameters of the mixture model, we used the (generalized inverse ) And the Stepwise Regression procedure

Regression models are one of the most important models used in modern studies, especially research and health studies because of the important results they achieve. Two regression models were used: Poisson Regression Model and Conway-Max Well-  Poisson), where this study aimed to make a comparison between the two models and choose the best one between them using the simulation method and at different sample sizes (n = 25,50,100) and with repetitions (r = 1000). The Matlab program was adopted.) to conduct a simulation experiment, where the results showed the superiority of the Poisson model through the mean square error criterion (MSE) and also through the Akaiki criterion (AIC) for the same distribution.

Paper type:

Abstract

The Classical Normal Linear Regression Model Based on Several hypotheses, one of them is Heteroscedasticity as it is known that the wing of least squares method (OLS), under the existence of these two problems make the estimators, lose their desirable properties, in addition the statistical inference becomes unaccepted table. According that we put tow alternative,  the first one is  (Generalized Least Square) Which is denoted by (GLS), and the second alternative is to (Robust covariance matrix estimation) the estimated parameters method(OLS), and that the way (GLS) method neat and certified, if the capabilities (Efficient) and the statistical inference Thread on the basis of an acceptable