摘要
本文提出一种马尔可夫交换人工神经网络,应用于经济市场中的黄金市场的波动性建模与预测。本文所提出的模型在条件波动过程的动态性与传统神经网络模型相比,在预测能力上有所不同。在本文中,应用此类模型来检验黄金收益率的波动性。对绝对误差、均方误差和均方根误差准则加以评估,并且在相同精度下进行改良的Diebold Mariano测试。为黄金市场日收益的预测提供了一个实证应用,结果表明,该方法在模拟和预测国际黄金日收益波动性方面具有较好的效果。
This paper presents a Markov switching artificial neural network(HMM),which is applied to the volatility modeling and prediction of the gold market in the economic market. The proposed models differ in terms of both the dynamics of the conditional volatility process and the forecasting capabilities compared to traditional neural network models. In this paper,these models are used to test volatility of gold return. MAE,MSE and RMSE are evaluated and the improved Diebold Mariano test is carried out under the same precision. An empirical application is provided for forecasting daily returns in gold market. The results suggest that the proposed approach performs well in modeling and forecasting volatility in daily returns of international gold market.
作者
谢荣燕
XIE Rong-yan(Business School,Hohai University,Nanjing 211100,China)
出处
《计算机与现代化》
2018年第6期12-15,共4页
Computer and Modernization
关键词
人工智能
神经网络
经济市场
波动率
马尔可夫
artificial intelligence
neural network
economic market
volatility
Markov