摘要
基于Gauss过程机器学习算法,通过分析股票样本的历史数据噪声问题,给出相应的股票样本数据回归预测模型,解决了股票异常数据的检测问题;并用蚁群算法,解决了Gauss过程机器学习算法的参数自适应问题.实验结果表明,该算法与其他算法相比,可在保证近似准确性的基础上,大幅度提高计算效率,提升用户满意度.
On the basis of the analysis of historical data on the stock sample, we proposed an algorithm for the prediction of stock data to find the abnormal data to detect the abnormal data, with the introduction of Gaussian process machine learning method. The adaptive mechanism for the parameters of the Gaussian process was also solved with ant colony algorithm. Finally, some experiments show that the proposed algorithm can improve the accuracy and enhance customers satisfaction.
出处
《吉林大学学报（理学版）》
CAS
CSCD
北大核心
2012年第6期1228-1232,共5页
Journal of Jilin University:Science Edition
基金
吉林省发改委高新技术项目(批准号:20106421)
吉林省重点科技发展计划项目(批准号:20100309)
吉林省教育厅科研项目(批准号:2012184)
关键词
异常数据
GAUSS过程
机器学习
蚁群算法
baseline algorithm
Gaussian process
machine learning
ant colony algorithm
