Construction of Scholar Portraits and Intelligent Analysis Method for Academic Information Based on Pre-Trained Models

The aim of this paper is to propose an intelligent analysis method based on pretrained models for academic information, which can be used to construct scholar profiles, explore research hotspots, and mine scholar collaboration networks. By integrating academic papers, patents, scientific research projects and activity data of social media, the Pretrained(such as ChatGLM) model is used to construct a multidimensional scholar portrait, including the natural attributes of scholars, academic achievements, cooperation networks and research interests. Based on the generated scholar profiles, we analyze the evolution of research interests of scholars and construct a scholar collaboration network with the aim of dynamically tracking changes in research directions and revealing academic collaboration relationships. The experimental results show that the performance evaluation demonstrates good performance on key metrics such as precision, recall, and F1-score, proving the potential of the proposed method in academic management and research resource allocation. This paper provides intelligent technical support for smart education and research management and offers a data-driven solution for predicting future research trends.