基于特征融合的虚假招聘广告检测模型  

Feature Fusion-Based Model for Fake Job Advertisement Detection

作　　者：王永政 纪淑娟 WANG Yongzheng;JI Shujuan(School of Computer Science and Engineering,Shandong University of Science and Technology,Qingdao 266590,China)

机构地区：[1]山东科技大学计算机科学与工程学院,山东青岛266590

出　　处：《软件导刊》2025年第2期26-32,共7页Software Guide

基　　金：国家自然科学基金项目(71772107)。

摘　　要：网络在线招聘逐渐取代了传统线下招聘,成为求职者首选的求职方式,但虚假招聘广告的出现给企业与求职者带来极大困扰,严重妨碍了网络招聘健康发展。针对现有基于单一机器学习模型的检测准确性较低,基于深度学习模型的时间效率较差的问题,提出一种基于特征融合的虚假招聘广告检测模型。首先,引入注意力机制为每个基分类器分配权重;其次,横向融合多个基分类器特征以显著提升虚假招聘广告检测效果。实验表明,特征融合模型准确率相较于性能最好的机器学习模型、BERT深度学习模型、TextCNN深度学习模型分别提升1.78%、1.67%、1.16%,模型运行时间略高于各机器学习模型但远低于深度学习模型,在EMSCAD数据集上的实验表明特征融合模型具有较好的虚假招聘广告检测性能。Online recruitment has gradually replaced traditional offline recruitment and has become the preferred way for job seekers,but the emergence of false recruitment advertisements has brought great trouble to enterprises and job seekers,and seriously hindered the healthy development of online recruitment.In order to solve the problems of low detection accuracy and poor time efficiency of existing single machine learning models based on deep learning models,a false recruitment advertisement detection model based on feature fusion was proposed.Firstly,the attention mechanism is introduced to assign weights to each base classifier.Secondly,multiple base classifier features were horizontally fused to significantly improve the detection effect of fake recruitment advertisements.Experiments show that the accuracy of the feature fusion model is 1.78%,1.67% and 1.16% higher than that of the machine learning model,BERT deep learning model and TextCNN deep learning model,respectively,and the running time of the model is slightly higher than that of the machine learning models but much lower than that of the deep learning model.Experiments on the EMSCAD dataset show that the feature fusion model has good performance in detecting false recruitment advertisements.

关 键 词：机器学习 特征融合 注意力机制 虚假招聘广告检测 深度学习 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]