ABSTRACT---- Replication of observation allows consistent estimation of slope parameter of a linear structural model when the ratio of variances is unknown or when some external information about parameters is not available. In this paper, we look at the way a linear structural relationship model work by replicating observations with two different estimation methods of slope parameter and different cases of existence of outliers. The maximum likelihood estimate (MLE) and a new nonparametric robust estimation method are used to estimate the slope parameter in replicated linear structural relationship model (RLSRM). The simulation studies and the application of real data are used to investigate the performance of the estimated parameters. Keywords— Maximum likelihood method, A nonparametric method, Trimmed mean, Outlier, Linear structural relationship model with replicated.
AMEL SAAD AB ALSHARGAWI(1-2022)

لهذه الدراسة هدفان مهمان وهما: أولاً: توضيح فكرة بناء الشبكات؛ العصبية المقترحة ثانياً: مقارنة هذه الطرق بالإدراك الجيد لنماذج السلاسل الزمنية (ARIMA) باستعمال المعيار MSE، وهو المعيار الأول لتدريب الشبكة العصبية والثاني لحساب آلية توقعات نماذج الشبكات العصبية. باستخدام بعض الأمثلة الخاصة اتضح أن الإجراءات حول نموذج الشبكات العصبية وجدت بأنها تقدم توقعات أفضل من نماذج السلاسل الزمنية، وأن نماذج الشبكات العصبية قد تستعمل في التنبؤ ببيانات السلاسل الزمنية بتعديل بعض الأوزان التى تعتبر معالم نماذج الشبكات العصبية والتى يمكن أن تقدر خلال عملية تدريب الشبكة، ودقة التوقعات مقدرة بالدالة المناسبة التى تستعمل في عملية تدريب الشبكة. إن مشكلة تنبؤ النماذج شائعة في التحليلات الإحصائية، وفى الغالب الطرق مستعملة للتعامل مع تنبؤ نموذج الانحدار والسلاسل الزمنية بالرغم من أن هذه الطرق قد لاتكون دقيقة في العينات الصغيرة و النتائج المتحصل عليها في هذا البحث حسبت بفصل مجموعة البيانات إلى مجموعتين جزئيتين أو أكثر، استعملنا الجزء الأول لملائمة النموذج والجزء الأخير لبناء التوقع باستخدام المعيار MSE كأداة للمقارنة بين النماذج, وكلما كانت قيمة هذا المعيار صغيرة كان النموذج أفضل. Abstract This study has two objectives. First, presenting artificial neural networks (ANN) second, comparing the proposed method with the well known ARIMA model, the accuracy of the neural network forecasts is compared with the corresponding ARIMA models by using the mean square error (MSE). By using the proposed (MSE) measures the artificial neural networks (ANN) were found deliver a better forecasts than the ARIMA model. A class of artificial neural networks (ANN) may be used in forecasting time series data. It may be used to approximate unknown expectation function of future observation given past values , thus the weights of these ANN can be viewed as parameters, which can be estimated through the network training. Then the model is used for forecasting. The accuracy of the forecasts is evaluated by suitable function. The problem of forecasting model is common in statistical analysis. One of the mostly used approach to deal with forecasting model is regression and time series. Although, approaches may not accurate in small sample. In an effort to forecast daily flow waters to the three important dams such as Ejdabia, Sirt, Benghazi, we will training to a take new tool if forecasting model which known as neural network model. This tool deal with testing data after made as partition of the original series into two sets first is called training set, were used to fit the model, while the second is called testing sets, were used to make forecasting. In this work the MSE is well known as tool for comparing between the models, further more when the MSE is less, the value of this model is a better than other models.
ساميه محمد ميره (2010)

In this study, the linear structural relationship model’s slope parameter is determined by using the proposed robust nonparametric method based on trimmed mean. This method is an upgrade to the nonparametric method that was put forward by Al-Nasser and Ebrahem (2005) by employing trimmed mean for all likely paired slopes rather than median slopes. Simulation study and real data were used to compare the proposed method’s performance versus the traditional maximum likelihood method. In the simulation study, based on both methods’ mean square error, it was inferred that the MLE method breaks down due to the presence of outliers even though its functioning was not affected when there was no outlier in the data set. Based on the real life example, it can be concluded that the performance of our proposed method was quite well in determining slope parameter
Amel Saad Alshargawi, (1-2022)