الفهرس 
NEURAL NETWORK AND ROUGH SETSOnly 14 pages are availabe for public view
 |  | from | 135 |  |  |
| | | from | 135 | | |
Abstract
Soft computing methodologies are widely used for solving real life problems as they provide a flexible information processing capability for handling ambiguity, uncertainty and incompleteness prevalent in real life data. Because soft computing systems have limitations, there is a demand for complex systems combining various approaches. Generally, Artificial Intelligence (AI) systems can be combined with other AI systems or with mathematical tools. In this thesis we design a Hybrid Rough Neural Prediction (HRNP) model, where rough set method is applied as a tool for reducing and choosing the most relevant sets of internal states for prediction. It acts as a preprocessing machine to reduce the features with minimal factors sets generated. The case study which we use is the problem of predicting the buy/sell points of Egypt stock price index. depending on the historical data of the index movement in the past. The target of this model is to get the state of the price index in the next week. The directions are categorized as 0, 0.5 and 1 in the data. A class value of 0 means fall in the stock, and a class value of 1 means rise in the stock price. A class value of 0.5 means that there is no change in the price between the present week and the previous week. This thesis is organized as follows:
Chapter 1: gives an introduction to neural network with its architectures and its characteristics. It also gives an introduction to rough sets with its benefits and its characteristics.
Chapter 2: illustrates prediction in general and stock market prediction as a special case with a review of some previous studies on stock market prediction.
Chapter 3: states our problem and discusses the proposed Hybrid Rough Neural Prediction (HRNP) system and its methodology in details.
Chapter 4: discusses our case study that is the problem of determining the trend of the Egypt Stock Market Price Index, and implementing of Hybrid Rough Neural Prediction model algorithm.
chapter 5: contains a conclusion of our thesis and the future work that we can do depending on this thesis.