استخدام الانحدار الرتيب لتقدير متوسط الإنفاق الشهري للأسرة العراقية لسنة 2005

الملخص EN

The data for this research was obtained from the annual statistical issued from central statistical organization which includes average monthly household expenditure on totals of goods and services and the average monthly household income for the year (2005) of (18) category of the monthly family income categories.

In this research, a nonparametric method called (Monotonic Regression) is used, it is a powerful and flexible method to design monotonic data.

we use increasing data and classify it into blocks according to the behavior of the data, and for each block we will get a straight line with its parameters.

After that we will obtain the estimation values for the response variable from these parameters inside each block.

The methods were used in this research (Isotonic Regression Estimator), (Daniel) test, these were used to test the trend, and we implement the (Backward Elimination Procedure) to see if there is a difference between isotonic regression and reduced isotonic regression.

This procedure is reduced the number of level sets by combining those whose values do not differ greatly.

some of the resulting solution blocks could be pooled together especially these with the estimated values do not differ a lot from their neighboring blocks, either solution blocks do not improve significantly, the fit can be eliminated, a model with fewer solution blocks fits better.

After combining the blocks, through finding a statistical (Am), and compare it with the value of to specify the appropriate number of level set, till we reach to (Am) is grater than (F), we will stop combining blocks, and so we arrived to isotonic regression differ from reduced isotonic regression, as well as the number of blocks has decreased, and the model could be better for applying.

And finding the estimation of the response variable (Y).

The aim of this research is to use a non-parametric statistical method in estimating the average monthly expenditure of the Iraqi family for the year 2005.

And in the case of the problem of the failure of meet the requirements of the parameterized method of testing and evaluation, our committees have a more flexible and manageable approach to data called non-parametric methods, including the monotonic regression in this research.

And the researcher's conclusion is that the average monthly household expenditure can be estimated by using non-parametric methods because of the high flexibility enjoyed by non-parametric capabilities compared to the parametric capabilities by measuring the parameter capacity.