In the non-stationary financial market, considering that earlier observations may have little or no relevance to the current investment decision making, we design two kinds of adaptive online portfolio strategies only based on recent historical data. Firstly, we design an adaptive online portfolio strategy by linearly combining the last portfolio and the best constant rebalanced portfolio corresponding to the recent historical data, which we call moving window. We determine the length of the moving window by adaptive learning. More precisely, we consider the strategies that always adopt the best constant rebalanced portfolio corresponding to the moving window of different fixed lengths as different experts, and at the beginning of the current period, we choose the length of moving window the same as the expert achieving maximum current cumulative return. Furthermore, we determine the length of moving window by only using the recent historical data to adaptively learn, and design another adaptive online portfolio strategy. We present numerical analysis by using real stock data from the US and Chinese markets, and the results illustrate that our strategies perform well, compared with some benchmark strategies and existing online portfolio strategies.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.