基于行为特征及关系网的大学生就业推荐算法研究

Research on employment recommendation algorithm for college students based on behavioral characteristics and relationship networks

大学生行为特征和企业特征为就业推荐算法提供了建模依据,其实现方式为采集数据、提取特征、训练算法模型。数据采集阶段需进行分类和预处理,确保数据格式、数值等符合要求。文章依托深度神经网络,提取了大学生行为序列特征,建立了PRHN推荐算法的理论模型。关系网可充分发掘学生数据和企业数据的图谱结构,有助于提升PRHN就业推荐算法的命中率,改善推荐集内的企业排序。因而可将关系网融入推荐算法,辅助完成推荐任务,提升算法性能和效果。

College students’behavior characteristics and enterprise characteristics for the employment recommendation algorithm to provide a modeling basis,the implementation is to collect data,extract features and train algorithm model.In the stage of data acquisition,classification and preprocessing are needed to ensure that the data format and numerical value meet the requirements.The feature of college students’behavior sequence is extracted by depth neural network,and the theoretical model of PRHN recommendation algorithm is established.The network can fully explore the graph structure of student data and enterprise data,which helps to improve the hit rate of PRHN’s algorithm and the ranking of enterprises in the recommendation set.Therefore,the network can be incorporated into the recommendation algorithm to assist the completion of the recommendation task and improve the performance and effectiveness of the algorithm.