Topic space trajectories

Abstract The annual number of publications at scientific venues, for example, conferences and journals, is growing quickly. Hence, even researchers it becomes harder to keep track research topics their progress. In this task, can be supported by automated publication analysis. Yet, many such methods result in uninterpretable, purely numerical representations. As an attempt support human analysts, we present topic space trajectories , a structure that allows the comprehensible tracking topics. We demonstrate how these interpreted based on eight different analysis approaches. To obtain results, employ non-negative matrix factorization as well suitable visualization techniques. show applicability our approach corpus spanning 50 years machine learning from 32 venues. addition thorough introduction method, focus extensive results achieved. Our novel method may employed paper classification, prediction future topics, recommendation fitting journals submitting unpublished work. An advantage applications over previous lies good interpretability obtained through methods.