Aggregating closing position experts for online portfolio selection

Online portfolio selection is a decision-making process that involves dynamically adjusting asset positions based on historical price sequence. Existing online portfolio strategies consecutively invest in risky assets regardless of market environment. When market situations are poor, the returns may suffer losses. In this paper, we propose a novel online portfolio strategy based on closing position signal. First, we obtain the signal of closing position based on the average value of relative prices using the exponential moving average method, and construct an investment decision. Second, we apply the online gradient update algorithm to aggregate expert strategies and propose our strategy. Then, we prove that the regret of the strategy has a theoretical upper bound. Finally, we conduct numerical experiments using real financial data from different markets. The results show that our strategy has good competitive performance in terms of cumulative wealth and risk-adjusted returns.