AIQF Daily digest — 8 new papers

Quant Paper Radar · 3 of 8

1. 1

   Accurate stock movement prediction via centroid-based randomness smoothing

   Yejun Soun, Hosung Lee, U Kang · OPENALEX · Machine-learning stock forecasting

   quant relevance 0.95

   Craft predicts next-day stock direction from historical prices and reports an investment simulation with Sharpe-ratio and drawdown results. Its price-only method can be tested on US equities, but the platform cannot reproduce its non-US evaluations or an exact historical S&P 500 membership universe.

   original paper

2. 2

   Entropy Regularization under No-Shorting: Structure Preservation and Convexity in Dynamic Mean-Variance Portfolio Selection

   Xinagyu Cui, Xun Li, Ziyan Li et al. · ARXIV · Dynamic Portfolio Optimization

   quant relevance 0.94

   The paper develops a long-only dynamic mean–variance allocation method with entropy regularization and a temperature restriction needed to preserve convexity. Its rolling application uses Fama–French industry portfolios; the described empirical finding concerns the admissible temperature region, not a stated trading-performance result.

   original paper

3. 3

   Three options are roughly enough - Variance-optimal semi-static hedging in the rough Heston model

   Paolo Di Tella, Martin Keller-Ressel, Beatrice Ongarato · ARXIV · Derivatives hedging

   quant relevance 0.92

   The paper develops a semi-static hedge for variance swaps using dynamic trading in the underlying and static positions in a small basket of European options. Its reported reduction in hedging error comes from numerical analysis of the rough Heston model, not a historical trading backtest.

   original paper

5 more papers written up in this period. Open the Quant Paper Radar →