<ul><li>Fund allocation in the financial domain is a crucial issue, with traditional methods facing challenges such as goal mismatch and uncertainty in time series forecasting.</li><li>To address these challenges, a Risk-aware Time-Series Predict-and-Allocate (RTS-PnO) framework is introduced, offering features like end-to-end training, forecasting uncertainty calibration, and model agnosticism.</li><li>RTS-PnO is evaluated on offline experiments using financial datasets and an online experiment in Cross-Border Payment business, showcasing superior performance over competitive baselines.</li><li>The code for the offline experiment is accessible at https://github.com/fuyuanlyu/RTS-PnO for further reference.</li></ul>

Timing is important: Risk-aware Fund Allocation based on Time-Series Forecasting

Discover more