摘要
从股票投资预测的技术发展方向来看,不同人工智能学习算法之间的组合学习日益得到关注。基于组合学习的思想,提出了股票投资短期预测的多模匹配识别算法(MPMA)。算法通过迭代计算数据采样频率、聚类分组、模式匹配将股价预测和涨跌预测纳入到一个统一的学习框架中,建立起不同人工智能学习算法之间的组合学习模型。实验结果表明,所提算法具有较好的预报和泛化能力。
In short-term prediction of stock investment, ensemble learning algorithms of artificial intelligence have been paid more and more attention. Based on the idea of ensemble learning, Multi-Pattern Matching Algorithm (MPMA) for short- term prediction of stock investment was proposed. The algorithm incorporates the stock price forecasting and trend prediction into a unified learning framework based on iterative computation of sampling frequency, clustering, pattern matching, and establishes an ensemble learning model among different artificial intelligence algorithms. The experimental results show that the proposed algorithm has good prediction and generalization ability.
出处
《计算机应用》
CSCD
北大核心
2014年第A02期180-183,共4页
journal of Computer Applications
基金
齐鲁证券校企合作研究基金资助项目(Y24101J1G2)
关键词
股票投资
短期预测
人工智能
聚类
分类
回归
stock investment
short-term prediction
artificial intelligence
clustering
classification
regression