期刊名称：Journal of Systemics, Cybernetics and Informatics
印刷版ISSN：1690-4532
电子版ISSN：1690-4524
出版年度：2017
卷号：15
期号：5
页码：74-80
出版社：International Institute of Informatics and Cybernetics
摘要：Volatility is a measurement of the risk of financial products. Astock will hit new highs and lows over time and if these highsand lows fluctuate wildly, then it is considered a high volatilestock. Such a stock is considered riskier than a stock whosevolatility is low. Although highly volatile stocks are riskier, thereturns that they generate for investors can be quite high. Ofcourse, with a riskier stock also comes the chance of losingmoney and yielding negative returns. In this project, we will usehistoric stock data to help us forecast volatility. Since thefinancial industry usually uses S&P 500 as the indicator of themarket, we will use S&P 500 as a benchmark to compute the risk.We will also use artificial neural networks as a tool to predictvolatilities for a specific time frame that will be set when weconfigure this neural network. There have been reports thatneural networks with different numbers of layers and differentnumbers of hidden nodes may generate varying results. In fact,we may be able to find the best configuration of a neural networkto compute volatilities. We will implement this system using theparallel approach. The system can be used as a tool for investorsto allocating and hedging assets.
