IRAQI JOURNAL OF STATISTICAL SCIENCES

In this paper we have investigated three algorithms. In the first algorithm we have derived a new optimal step size gradient algorithm which is preferable over the classical SD algorithm both in theory and in the real computation. In the second algorithm we have derived and implemented a new formula for the non-quadratic model with a new . In the third algorithm we have tried to make a new hybrid algorithm between the above three different step sizes.
Our numerical results are promising in general by implementing ten non-linear different test functions with different dimensions.

In this Research, we employ Principal Components manner for the forecasting in the time series , we are makig the forecast to a data which used from the researcher ( Ljung,1999) , when we used the principal components,we obtain a good forecasting with a good statistcal specifics

In this paper we present a technique for computing the minimum value of an objective function in the frame of gradient descent methods based on combination of Barzilai and Borwein approximation of Hessian matrix of objective function and Lipchetz constant in the gradient flow algorithm which is derived from a system of ordinary differential equations associated to unconstrained optimization problem. This algorithm suitable for large- scale unconstrained optimization problems, computational results for this algorithm is given and compared with BB method showing a considerable improvement.

The importance of the classification of the satellite images appeared as one of the important subjects which the researchers were interested in, the traditional and hybrid methods were employed in the research.
In the traditional method the decision tree and K-means clustering algorithm were used, which gave bad and accepted results in classifying satellite images, so the importance of genetic hybrid algorithm study is concerned with classifying the satellite images.
Also a hybrid system was suggested between the genetic algorithm and the k-means clustering algorithm, which gave good results in classifying the satellite images.

This paper deals with the comparison between some linear filter for time series represented by Weiner filter through autoregressive model,moving averagemodel and difference filter. An application for lung cancer Time Series in Mosul city during (1990-2000) are also considered.

This study finds out a new algorithm for scheduling problem in computer, This algorithm was compared with HLFET(Highest Level First With Estimated Times Scheduling Algorithm) and SCFET (Smallest Co-Levels First With Estimated Times Scheduling Algorithm) scheduling algorithms.
By letting the communication cost is equal to zero, It shoues high efficiency and easy to use.

This research concentrates on the study of genetic algorithm which is regarded as one of the approaches of modern artificial intelligence an algorithm is suggested to transfer a gene to solve the problems of transportation. The application of this algorithm to the matters of balanced transportation led to a number of solutions equal to the number of generation including the most ideal solution (the least possible total cost). In addition, a new leap is suggested through the algorithm of genetic transference whose use led to tangible importance in the results (reducing the value of the total cost).

In factor analysis, whenever the number of variables (m) less than the number of the experimental units (n) the procedure of R-mode should be applied, if (m) exceed (n) the Q-mode is the suitable procedure to follow. A shortcut transformation between the results of the two procedures were established in any analytical step, they are: eigen vectors, factor loadings, factor coefficients and factor scores.

This paper studied the relationship for some variables that direct and indirect effect for the agriculture investment and agriculture local output for the period (1980-2000), to help us to identification how this variables affect to the agriculture sector and what extent responding for this sector to this variables, for estimation we apply (Full Information Maximum Likelihood) (FIML) to estimate the parameters by using (SAS) and (MINITAB) Packages.

In this research, we use on ( ) and ( ) two criterions to compare between the estimators of Adaptive Regression model and the estimators of Ordinary and General Least Squares, we found that, Adaptive Regression estimator is the best by comparing with estimators because minimum mean square error.