About Our Project

Research trend analysis provides an important starting point for understanding how scientific fields are structured and how they evolve over time. Scientific publications have grown rapidly (Figure 1), and when viewed across years and decades, they reveal shifting patterns of research focus and scientific interest. As millions of papers are published each year now, there is an increasing need for analytical methods that can characterize research fields from both structural and temporal perspectives. Such approaches can help researchers track the evolution of scientific interests, identify emerging directions, and uncover underexplored areas with potential for future discovery.

PubMed is one of the largest and most widely used databases for biomedical literature, providing a comprehensive and time-resolved record of scientific publications. With more than 40 million citations and abstracts, it covers a broad spectrum of biomedical research.

Recent advances in large language models (LLMs) have made it increasingly feasible to process large-scale text corpora and extract semantic information from unstructured scientific literature. These developments provide new opportunities to analyze publication data at a scale and depth that were previously difficult to achieve.