Processing Outlier values in time series using Sectional data models
A thesis
Submitted to the council of the college of administration and Economic\ University of Karbala, as partial fulfillment of the requirements for the degree of Master of Science in statistics
By
Montazer Mustafa Nassif
Super vised by
Asist.Prof. Dr. Jassim N. Hussain

Abstract
The presence of outliers values in the number of people infected with THE Coronavirus (Coved-19) for August 2020 in Iraq affects the results of statistical analysis of these data, and these data can be obtained through repeated observations of a phenomenon about N of cross-sections during a certain Time series , The letter referred to the analysis of data related to the coronavirus pandemic (COVED-19) before diagnosing and treating abnormal observations using ct data models (fixed effect model and random impact models), and then addressed in this study the diagnosis of abnormal observations in these data in ways (Drawing the box, the hat matrix, the deleted Stuart protector), and then the ways of addressing these abnormal views by methods (deletion, log conversions, trimming abnormal views), and the data was analyzed again The results on the applied side showed the preference of the fixed effect model before and after the treatment of abnormal observations of the random impact model through the use of houseman test and the test of the selection coefficient (R2) and the overall moral test of model (F), and the results also showed that the treatment of abnormal values did not improve and estimate the efficiency of the models used, and the results of the Durbin-Watson test (D.W) showed that the models used suffered from a positive subjective association between random errors. After processing abnormal observations using ( Fixed effect model and random impact model.