Please use this identifier to cite or link to this item:
|Title||Comparative Approach of Artificial Neural Network and ARIMA Models on Births per Month in Gaza Strip using R|
|Title in Arabic||دراسة مقارنة لنماذج الشبكة العصبية الاصطناعية و نماذج اريما على مواليد قطاع غزة باستخدام برنامج r|
Time series modeling and forecasting has fundamental importance to various practical domains. Thus a lot of active research works is going on in this subject during several years. Many important models have been proposed for improving the accuracy and efficiency for time series modelling and forecasting. In this thesis we will build models to predict the number of Births in Gaza Strip, based on the number of births registered in the Palestinian Health Information Center (PHIC) at the Ministry of Health in Gaza Strip over the years, which were recorded in the time period from January 2000 to December 2013. Study models on the number of Births registered in the period from 2000 to 2012 and containing 156 observation, and we will keep the number of Births for the year 2013 to use it in comparing the predicted values from the models and ensure the accuracy of the prediction. An approach of artificial neural networks to number of Births per month in Gaza Strip is presented in this work. Four different architectures of artificial neural networks have been trained and tested to forecast monthly Births in Gaza Strip. The performance of two feedforward Neural Networks (multilayer perceptron and radial basis function), and two recurrent (Elman and Jordan) Neural Networks will be analysed, and Compared with the expected pattern of Box Jenkins model. To evaluate forecast accuracy as well as to compare among Box-Jenkins and four ANNs models fitted to a time series, we have used the four performance measures, MSE, MAE, RMSE, and MAPE, the model has lowest value of them is superior to the other models. In this context, we found that MLP model outperforms SARIMA, Jordan recuurent neural netwotk model (JRNN) out performs SARIMA and MLP model , Elman Recurrentm neural network (ERNN) model outperforms JRNN model, and Radial Bais feed forward neural model (RBFNN ) outperforms JRNN.
|Publisher||الجامعة الإسلامية - غزة|
|Files in this item|