为实现对非平稳、非线性股票价格时间序列的高精度预测,提出经验模态分解下基于支持向量回归的股票价格集成预测方法EMD-SVRF(EMD and SVR based stock price integrated forecasting)。首先,运用经验模态分解方法获得股票对数收益率时...为实现对非平稳、非线性股票价格时间序列的高精度预测,提出经验模态分解下基于支持向量回归的股票价格集成预测方法EMD-SVRF(EMD and SVR based stock price integrated forecasting)。首先,运用经验模态分解方法获得股票对数收益率时间序列的本征模函数及趋势序列,然后,利用ε不敏感支持向量回归为各本征模函数及趋势序列分别建立预测模型,并计算各本征模函数及趋势项的预测值,最后,集成得到股票收益率序列预测值。实验表明,相对现有的EMD-Elman网络和ARMA-GARCH等主流股价预测方法,EMD-SVRF具有更小的拟合误差和预测误差,是一种高精度的股票价格预测方法。展开更多
The attack graph methodology can be used to identify the potential attack paths that an attack can propagate. A risk assessment model based on Bayesian attack graph is presented in this paper. Firstly, attack graphs a...The attack graph methodology can be used to identify the potential attack paths that an attack can propagate. A risk assessment model based on Bayesian attack graph is presented in this paper. Firstly, attack graphs are generated by the MULVAL（Multi-host, Multistage Vulnerability Analysis） tool according to sufficient information of vulnerabilities, network configurations and host connectivity on networks. Secondly, the probabilistic attack graph is established according to the causal relationships among sophisticated multi-stage attacks by using Bayesian Networks. The probability of successful exploits is calculated by combining index of the Common Vulnerability Scoring System, and the static security risk is assessed by applying local conditional probability distribution tables of the attribute nodes. Finally, the overall security risk in a small network scenario is assessed. Experimental results demonstrate our work can deduce attack intention and potential attack paths effectively, and provide effective guidance on how to choose the optimal security hardening strategy.展开更多
文摘为实现对非平稳、非线性股票价格时间序列的高精度预测,提出经验模态分解下基于支持向量回归的股票价格集成预测方法EMD-SVRF(EMD and SVR based stock price integrated forecasting)。首先,运用经验模态分解方法获得股票对数收益率时间序列的本征模函数及趋势序列,然后,利用ε不敏感支持向量回归为各本征模函数及趋势序列分别建立预测模型,并计算各本征模函数及趋势项的预测值,最后,集成得到股票收益率序列预测值。实验表明,相对现有的EMD-Elman网络和ARMA-GARCH等主流股价预测方法,EMD-SVRF具有更小的拟合误差和预测误差,是一种高精度的股票价格预测方法。
基金Supported by the National Natural Science Foundation of China(61373176)the Natural Science Foundation of Shaanxi Province of China(2015JQ7278)the Scientific Research Plan Projects of Shaanxi Educational Committee(17JK0304,14JK1693)
文摘The attack graph methodology can be used to identify the potential attack paths that an attack can propagate. A risk assessment model based on Bayesian attack graph is presented in this paper. Firstly, attack graphs are generated by the MULVAL（Multi-host, Multistage Vulnerability Analysis） tool according to sufficient information of vulnerabilities, network configurations and host connectivity on networks. Secondly, the probabilistic attack graph is established according to the causal relationships among sophisticated multi-stage attacks by using Bayesian Networks. The probability of successful exploits is calculated by combining index of the Common Vulnerability Scoring System, and the static security risk is assessed by applying local conditional probability distribution tables of the attribute nodes. Finally, the overall security risk in a small network scenario is assessed. Experimental results demonstrate our work can deduce attack intention and potential attack paths effectively, and provide effective guidance on how to choose the optimal security hardening strategy.
