The Center currently carries out several research projects, revolving around asset selection and allocation.
Behavioral Asset Allocation
Applying behavioral finance theory such as the cumulative prospect theory and regret theory to understand the behaviors of small investors, predict the impact of such behaviors on the market and mitigate behavioral risks.
Data Driven Robust Asset Allocation
Applying robust optimization techniques to construct data-driven portfolios without needing to estimate any parameters (especially the stocks expected return rates).
Dynamic Asset Allocation
Applying stochastic automatic control theory to asset allocation in order to achieve different objectives such as return—risk efficiency, index tracking or maximizing the probability of reaching a goal.
Investment Strategy Evolution
Applying machine learning techniques such as supervised learning and reinforcement learning to train and develop evolutionally superior investment strategies.
Portfolio Space Reduction
Applying network clustering technique based on correlations to dramatically reduce the number of assets in a portfolio while still maintaining a sufficient level of diversification.
Probability Weighting and Investment
Inferring from data the probability weighting, which is a behavioral anomaly that exaggerates the tails of an asset return distribution, and taking advantage of its impact on prices to optimize asset selection and allocation.
Reinforcement Learning in Finance
Applying Gaussian exploration in reinforcement learning to dynamic portfolio selection and devising model-free, data-driven algorithms to make investment decisions.