Academic Field and Future Influence Prediction for Scholar Profile Construction

Collecting scholar information from massive academic resources to construct scholar profiles can provide a reference for various academic activities. For a scholar profile, except basic attributes such as age, gender, job title, some potential attributes need to be predicted for describing the scholar. The scholars academic field and future influence are two important attributes for constructing profile. How to predict the potential attributes precisely is a challenge. Towards that, we propose two models to predict academic field and influence respectively for constructing scholar profile in this paper. In detail, we propose a double-layer attention model of paper information and journal information representation model for predicting scholars academic field based on multiple data sources. We also propose a two-level fusion model based on feature combination for predicting scholars’ future academic influence. Extensive experiments are conducted on AMiner dataset, and the results prove that our models are superior to the state-of-the-art methods.