AMA-K: Aggressive Multi-Temporal Allocation An Algorithm for Aggressive Online Portfolio Selection

TL;DR

AMA-K is an innovative algorithm for aggressive online portfolio selection that combines data mining, optimization, and machine learning to achieve higher returns with acceptable risk across diverse market conditions.

Contribution

It introduces a new Pattern-Matching methodology that enhances dynamic portfolio construction and outperforms existing strategies in key financial metrics.

Findings

Outperforms benchmarks in maximum drawdown, yield, and Sharpe ratio.

Achieves high returns with acceptable risk in various market conditions.

Demonstrates the effectiveness of integrating data mining and machine learning in portfolio strategies.

Abstract

Online portfolio selection is an integral componentof wealth management. The fundamental undertaking is tomaximise returns while minimising risk given investor con-straints. We aim to examine and improve modern strategiesto generate higher returns in a variety of market conditions.By integrating simple data mining, optimisation techniques andmachine learning procedures, we aim to generate aggressive andconsistent high yield portfolios. This leads to a new methodologyof Pattern-Matching that may yield further advances in dynamicand competitive portfolio construction. The resulting strategiesoutperform a variety of benchmarks, when compared using Max-imum Drawdown, Annualised Percentage Yield and AnnualisedSharpe Ratio, that make use of similar approaches. The proposedstrategy returns showcase acceptable risk with high reward thatperforms well in a variety of market conditions. We concludethat our algorithm provides an improvement in searching foroptimal portfolios compared to existing methods.