Reported out-of-sample Sharpe ratios above 4.6 for news-sentiment trading are likely inflated by using same-calendar-day GDELT events that include post-close news.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
q-fin.CP 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Interpretable Machine Learning for Macro Alpha: A News Sentiment Case Study
Reported out-of-sample Sharpe ratios above 4.6 for news-sentiment trading are likely inflated by using same-calendar-day GDELT events that include post-close news.