【Purpose/significance】This paper used text mining method to analyze the frontier topics and development trends of Univer⁃

sity papers, so as to provide a new perspective for university scientific research evaluation as well as a model reference and theoretical

support for university scientific research advantage direction selection.【Method/process】Firstly, the title information of high-level pa⁃

pers in Colleges and universities is obtained, the PLDA model is used to identify the research topics, and a recognition model is pro⁃

posed to represent the research frontier topics in colleges and universities by using the indexes such as topic heat and topic novelty.

Then, the cosine similarity model is used to calculate the research similarity between different topics, and the evolution of research top⁃

ics in colleges and universities is analyzed to reveal the development trend of research topics in colleges and universities. Through the

comparative analysis of the fit degree with the world research frontier content, the frontier areas of university research topics are identi⁃fied.【Result/conclusion】The experimental results show that the proposed method can effectively reveal the research frontier topics of universities, the evolution law of topics over time, and the fit degree with the world research frontier.【Innovation/limitation】This pa⁃per proposes a research method for the research frontier and evolution trend of university papers from the perspective of text topics.The main limitation is that the coverage of data sets is not comprehensive enough.