الفهرس يوجد فقط 14 صفحة متاحة للعرض العام
 |  | from | 176 |  |  |
| | | from | 176 | | |
المستخلص
Mining data streams has been at focus since last few years. It concerned with discovering valuable and hidden information and knowledge from continuous streams of data. Stock market prediction is considered one of the most commonly time series stream data applications. Stock market prediction is the process of attempting to determine the future value of a company stock based on its historical data. It has been at focus for years since the accurate prediction of a stock’s future price will maximize investor{u2019}s gains. The nature of stock market data is variable and nonlinear. Also stock time series behavior is close to random-walk. The selection of appropriate training and prediction methodology is still a very critical problem. It is difficult to find training algorithm that is suitable for all applications under all conditions. In this thesis, we proposed hybrid models that are based on financial technical indicators and are using natural inspired algorithms in training and optimizing of Least Square Support Vector Machine (LS-SVM) and Extreme Learning Machine (ELM) methods. The used algorithms are Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC), Bat Algorithm (BA), Modified Cuckoo Search (MCS), and Flower Pollination Algorithm (FPA)