Forecasting Research Trends by Mining Literature Findings
Alex Morales (University of Illinois at Urbana-Champaign)
Bringing unsolved long-standing research problems and new emerging topics to the attention of researchers in a timely manner can help accelerate research discovery as well as improve the productivity of researchers. With the vast emergence of research results as publications, the problem of information overload has become more prevalent in the academic community, making it difficult for researchers to keep track of all the new progress in research and identify unsolved key challenges, especially in interdisciplinary areas. In this work, we investigate research trends by mining these insights in academic literature. In particular we develop a probabilistic model to measure the degree of overlap between researcher’s insights and later related works. This model allows for identifying unexplored domains, long-standing problems, and new emerging topics.