الفهرس | Only 14 pages are availabe for public view |
Abstract Forecasting stock price of sales is an important factor for improving sales for retailers. When accuracy of stock price of sales is improved, this affects to the manufactures to increase more products and the company has more money. The investors are attracted to the company to sell or buy stocks because the company has more power in the market. This makes the manufactures distribute many products to the retailers and more sales are achieved to the organization such as Medical Union Pharmaceuticals (MUP) Company. The stock price of sales is an important factor for analyzing the sales for retailers and the organization. The main contribution of this thesis is analyzing the main factors (indicators) that affect on the stock price of sales. Forecasting the stock price of sales to Medical Union Pharmaceuticals (MUP) Company is developed using four techniques (Multi-layer perceptron (MLP), Radial Basis Function (RBF), forward stepwise and backward stepwise). A comparison between the forecasting results of these techniques is given. The results show that MLP is preferred than RBF because MLP achieved the minimum relative error with and without features reduction. It takes much shorter training time with and without features reduction than RBF. Forward stepwise regression and backward stepwise regression entered nine indicators and achieved coefficient of determination R square (R2) 0.956 respectively. |