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

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.