← Research

Master’s research · NLP & predictive analytics

News sentiment & market prediction

Explore relationships between news signals and historical stock-market movements.

NLP · Big Data · HPC

Master’s research at Southern Illinois University Carbondale · Study data, 2015–2017

The problem & approach

Compare machine-learning classifiers and ensemble methods using news and event signals from GDELT alongside daily Dow Jones Industrial Average prices.

Methods & contributions

  • Build predictive features from public news and event data.
  • Compare individual classifiers with ensemble methods.
  • Evaluate predictions using the historical data collected for the study.

Research connections