非平衡数据集下的高职学生就业预测模型  

Employment Forecasting Model of Higher Vocational StudentsBased on Unbalanced Data Set

作　　者：熊露露 年梅[1] 张俊 XIONG Lulu;NIAN Mei;ZHANG Jun(Xinjiang Normal University,Urumqi 830054;Xinjiang Railway Vocational and Technical College,Urumqi 830000;Xinjiang Institute of Physical and Chemical Technology,Chinese Academy of Sciences,Urumqi 830011)

机构地区：[1]新疆师范大学,乌鲁木齐830054 [2]新疆铁道职业技术学院,乌鲁木齐830000 [3]中国科学院新疆理化技术研究所,乌鲁木齐830011

出　　处：《计算机与数字工程》2023年第3期675-678,730,共5页Computer & Digital Engineering

基　　金：自治区高校科研项目(编号:XJEDU2017S032);新疆师范大学“数据安全”重点验室招标项目(编号:XJNUSYS102017B04)资助。

摘　　要：传统机器学习算法对不平衡数据进行二分类时,容易出现分类偏移问题,就业预测数据存在正负样本不平衡问题,为了提高就业预测的精度,论文设计了ADASYN-SMOTE-RF就业预测模型。首先使用ADASYN-SMOTE算法对训练集生成和扩充小类样本,然后使用随机森林(RF)算法建立预测模型。实验结果表明,ADASYN-SMOTE-RF模型较好地解决了样本不均衡导致的预测准确度不高的问题,为高职学生就业率的提高提供技术支持。When the traditional machine learning algorithm is used to classify the unbalanced data,it is easy to have the prob⁃lem of classification deviation.The employment forecast data has the problem of imbalance between positive and negative samples.In order to improve the accuracy of employment prediction,this paper designs the ADASYN-SMOTE-RF employment prediction model.Firstly,the ADASYN-SMOTE-RF algorithm is used to generate and expand the small class samples for the training set,and then the random forest(RF)algorithm is used to establish the model prediction model.The experimental results show that ADAAS⁃YN-SMOTE-RF model can solve the problem of low prediction accuracy caused by unbalanced samples,and provide technical sup⁃port for the improvement of employment rate of higher vocational students.

关 键 词：就业预测 ADASYN-SMOTE-RF 过采样处理 随机森林 就业率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]