A novel adjusted learning algorithm for online portfolio selection using peak price tracking approach

Dai, Hong-Liang; Huang, Cui-Yin; Dai, Hongming; Lai, F. Anthony; Lv, Xueshuai; Adnan, Rana Muhammad

doi:10.1016/j.dajour.2023.100256

Cited by 4 publications

(2 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…PP assumes the price continues to move at its maximum price in a recent time window but fails to capture the continuous depreciation of the asset price. Recently, there have been many works to handle the shortcomings of PP, for example, the trend peak price tracing [11] and the peak price tracking approach [18].…”

Section: Related Work and Problem Settingmentioning

confidence: 99%

Passive Aggressive Ensemble for Online Portfolio Selection

Xie,

Yin,

et al. 2024

Mathematics

View full text Add to dashboard Cite

Developing effective trend estimators is the main method to solve the online portfolio selection problem. Although the existing portfolio strategies have demonstrated good performance through the development of various trend estimators, it is still challenging to determine in advance which estimator will yield the maximum final cumulative wealth in online portfolio selection tasks. This paper studies an online ensemble approach for online portfolio selection by leveraging the strengths of multiple trend estimators. Specifically, a return-based loss function and a cross-entropy-based loss function are first designed to evaluate the adaptiveness of different trend estimators in a financial environment. On this basis, a passive aggressive ensemble model is proposed to weigh these trend estimators within a unit simplex according to their adaptiveness. Extensive experiments are conducted on benchmark datasets from various real-world stock markets to evaluate their performance. The results show that the proposed strategy achieves state-of-the-art performance, including efficiency and cumulative return.

show abstract

Section: Related Work and Problem Settingmentioning

confidence: 99%

Passive Aggressive Ensemble for Online Portfolio Selection

Xie,

Yin,

et al. 2024

Mathematics

View full text Add to dashboard Cite

show abstract

“…Compared to the previous PPT algorithm, TPPT boasts significantly improved price tracking capabilities. Dai et al [26] developed an adjusted PPT approach. By introducing specific parameters, the algorithm is able to fine-tune the influence of peak price and residual factor, resulting in a more accurate prediction of future prices.…”

Section: Ppt Algorithmmentioning

confidence: 99%

Online Adaptive Asset Tracking Algorithm with Ordinal Information

Lai,

Liang,

Huang

et al. 2023

JCCE

View full text Add to dashboard Cite

In recent years, an increasing number of researchers have applied machine learning techniques to online portfolio selection (OLPS), aiming to improve the efficiency and effectiveness of portfolio management in the digital field. In this study, we design and implement a novel OLPS algorithm called "online adaptive asset tracking algorithm" (OAAT). Compared to the peak price tracking (PPT) algorithm, it complements more historical information of assets in the investment portfolio and provides a more effective solution for parameter selection of the PPT algorithm. Firstly, the OAAT algorithm updates investment proportions by tracking the historical information of assets, which includes recent peak prices, historical returns, and historical volatility. Secondly, the OAAT algorithm optimizes parameters through online learning. The initial parameters are selected based on the minimum sum principle of the ordinal information. After each phase of trading, the parameters are optimized through the gradient descent algorithm, and the average values of the optimal parameters in the last 5 days are used as the parameters of the next phase. Finally, with the optimized parameters and the tracked asset information, the fast error backpropagation algorithm outputs the investment ratio through gradient projection. Compared with the benchmarks, follow-the-winner, follow-the-loser, and pattern-matching-based algorithms under four Hong Kong stock index constituents data sets and three US stock index constituents data sets, the empirical comparative analysis and statistical test show that the OAAT algorithm can effectively determine the investment proportion to balance return and risk.

show abstract