The Research on College Student Employment Recommendation Method by Integrating Temporal Text and Collaborative Filtering

In response to the limited application of recommendation systems in the education field, this paper focuses on the college student employment position recommendation and proposes a recommendation method that integrates temporal text and collaborative filtering. A time-decaying attention mechanism is introduced to model the temporal features of resume texts, capturing the dynamic trajectory of students’ project abilities. A cross-era comparison learning strategy is designed to alleviate the data sparsity issue. By adding a business rule injection layer, constraint functions are used to correct the text output based on rigid requirements such as education level and salary for job positions. Experiments on a real employment dataset evaluate the model’s performance, showing that both HR@N and recall rate results outperform baseline models. This provides valuable reference for the application of recommendation algorithms in the employment recommendation field.