基于递推最小二乘预测的被动进攻在线投资组合算法分析  

Passive aggressive online portfolio strategies based on recursive least squares forecasting

作　　者：李华[1] 何洪浏 LI Hua;HE Hong-liu(College of Science,Changchun University,Changchun 130022,China)

机构地区：[1]长春大学理学院,吉林长春130022

出　　处：《吉林师范大学学报（自然科学版）》2023年第2期53-60,共8页Journal of Jilin Normal University:Natural Science Edition

基　　金：国家自然科学基金项目(12171078);吉林省科技厅产业关键核心技术攻关课题(20220201160GX)。

摘　　要：在被动进攻算法(passive aggressive mean reversion, PAMR)的基础上引入了递推最小二乘,使用其预测值代替原先的相对价格,同时实证分析了国内外5个股票数据集.结果证明,递推最小二乘被动进攻算法均取得了更好的累计收益,证实了其更加优异的收益能力.Recursive least squares was introduced into the PAMR algorithm and its predicted values were used in place of the original relative prices,meanwhile,an empirical analysis of five domestic and international stock data sets was carried out.The recursive least squares passive aggressive algorithm achieves better cumulative returns,confirming its superior return capability.

关 键 词：被动进攻均值回归 递推最小二乘 在线投资组合 

分 类 号：F830.59[经济管理—金融学]