As an important part of improving the employment rate of college students, the current employment guidance work in colleges and universities suffers from problems such as lack of relevance and formality. In this paper, we start from the perspective of college students’ employment prediction, integrate the data mining method with college graduates’ employment prediction, and design the operation process of the employment prediction model. Then, we process the performance data and employment data of N college graduates separately, and use the Pearson correlation coefficient to analyze the correlation between graduates’ employment destinations and their demographic characteristics. Then the hybrid feature selection algorithm based on mutual information and weights (HMIGV) is introduced for the elimination of redundant information in the feature set. On this basis, the XGBoost algorithm is used to predict the employment of graduates, and a college student employment prediction model based on the HMIGV algorithm and the XGBoost algorithm is established. Compared with similar model algorithms, the model has the highest prediction efficiency on different college graduates’ datasets, with the lowest training time of 38.46 s and the lowest testing time of 6.82 s. By predicting college students’ employment with high efficiency, the model is able to provide powerful technical support and reference for graduates’ employment guidance work.
