摘要
对大学生所学课程进行分组和投影得到一组隐藏变量,这些隐藏变量之间的相关性很小,可以视为相互独立,将它们作为就业类别结点的子结点,由此得到一个三层次的树型网络结构,网络参数使用EM算法学习获得。通过和朴素贝叶斯形式的就业模型的对比发现,该模型就业预测的准确率高于朴素贝叶斯网,而且受就业样本数量的影响较小。
Grouping and projection courses learned by college students yields a set of hidden variables with little correlation.These hidden variables can be regarded as independent of each other,so can be taken as sub-nodes of employment category node,resulting in a three-level tree-based network structure.The parameters can be learned using the EM algorithm.The comparison with the employment model in the Naive-Bayesian form reveals that the employment prediction is higher than the Naive-Bayesian network and is less affected by the number of employment samples.
出处
《工业控制计算机》
2021年第11期108-110,共3页
Industrial Control Computer
基金
安徽省高校自然科学研究重点项目“贝叶斯网络在大学生就业能力培养中的应用研究”(KJ2019A1054)。
关键词
朴素贝叶斯网
就业能力
就业预测
Naive Bayes Network
employ ability
employment forecasting
