Portfolio Selection is an important real-world financial task and has attracted extensive attention in artificial intelligence communities. This task, however, has two main difficulties: (i) the non-stationary price series and complex asset correlations make the learning of feature representation very hard; (ii) the practicality principle in financial markets requires controlling both transaction and risk costs. Most existing methods adopt handcraft features and consider no constraints for the costs, which may make them perform unsatisfactorily and fail to control both costs in practice. In this paper, we propose a cost-sensitive portfolio selection method with deep reinforcement learning. Specifically, a novel two-stream portfolio policy network is devised to extract both price series patterns and asset correlations, while a new cost-sensitive reward function is developed to maximize the accumulated return and constrain both costs via reinforcement learning. We theoretically analyze the near-optimality of the proposed reward, which shows that the growth rate of the policy regarding this reward can approach the theoretical optimum. We also empirically evaluate the proposed method on real-world datasets. Promising results demonstrate the effectiveness and superiority of the proposed method in terms of profitability, cost-sensitivity and representation abilities.

To improve existing online portfolio selection strategies in the case of non-zero transaction costs, we propose a novel framework named Transaction Cost Optimization (TCO). The TCO framework incorporates the L1 norm of the difference between two consecutive allocations together with the principles of maximizing expected log return. We further solve the formulation via convex optimization, and obtain two closed-form portfolio update formulas, which follow the same principle as Proportional Portfolio Rebalancing (PPR) in industry. We empirically evaluate the proposed framework using four commonly used data-sets. Although these data-sets do not consider delisted firms and are thus subject to survival bias, empirical evaluations show that the proposed TCO framework may effectively handle reasonable transaction costs and improve existing strategies in the case of non-zero transaction costs.

We examine the performance of trend following strategies in Chinese commodity futures markets. We provide evidence that trend following‐based technical trading rules yield better performance than the buy and hold strategy on both individual contracts and sorted portfolios. The outperformance is robust to transaction costs, data frequency, sub‐prime crisis, shorting constraint, delayed execution, liquidity and parameters. Finally, the profitability of the trend following strategy may be subject to data snooping bias.

Machine learning and artificial intelligence techniques have been applied to construct online portfolio selection strategies recently. A popular and state-of-the-art family of strategies is to explore the reversion phenomenon through online learning algorithms and statistical prediction models. Despite gaining promising results on some benchmark datasets, these strategies often adopt a single model based on a selection criterion (e.g., breakdown point) for predicting future price. However, such model selection is often unstable and may cause unnecessarily high variability in the final estimation, leading to poor prediction performance in real datasets and thus non-optimal portfolios. To overcome the drawbacks, in this article, we propose to exploit the reversion phenomenon by using combination forecasting estimators and design a novel online portfolio selection strategy, named Combination Forecasting Reversion (CFR), which outputs optimal portfolios based on the improved reversion estimator. We further present two efficient CFR implementations based on online Newton step (ONS) and online gradient descent (OGD) algorithms, respectively, and theoretically analyze their regret bounds, which guarantee that the online CFR model performs as well as the best CFR model in hindsight. We evaluate the proposed algorithms on various real markets with extensive experiments. Empirical results show that CFR can effectively overcome the drawbacks of existing reversion strategies and achieve the state-of-the-art performance.

Studies consistently find that inflation is an important augmented factor for intertemporal capital asset pricing models (ICAPMs) when pricing the Fama–French 25 size and book-to-market portfolios. We extend this line of research by investigating two alternative ICAPM models (from Michel; Hahn and Lee) and the three-factor model from Hou et al. We find significant evidence that both ICAPMs and Hou et al.’s three-factor model perform better when augmented with inflation than the original models. The augmented models achieve a good model fit with the fewest factors, thus avoiding or alleviating the over-fitting problem.