基于LDA的招聘信息技能标签生成算法

Recruitment Information Skill Label Generation Algorithm Based on LDA

随着互联网的迅速发展,网络招聘成为主流招聘途径。大数据时代有价值的网络招聘信息往往隐藏在招聘文本中不易被发现,导致利用率不高。为了挖掘出招聘信息背后的价值,使招聘厂商对应聘者信息更方便地进行检索、分类和管理,提出基于招聘文本信息的技能标签生成算法。通过招聘词库提取关键词生成候选标签,然后将获得的文本信息进行LDA主题挖掘,对主题关键词与获取的候选标签进行相似度计算,最后对候选标签进行加权计算得出各主题下的技能标签。实验结果表明,该标签生成机制能反映出不同岗位的技能倾向,为招聘厂商合理高效地筛选应聘者提供参考依据。

With the rapid development of the Internet,the traditional mode of recruitment has been greatly impacted,and online recruitment has gradually become a mainstream way of recruitment.However,in the era of big data,valuable recruitment information is often hidden in the recruitment text,which is not easy to be found and thus cannot be effectively used.In order to dig out the value behind the recruitment information and better serve the recruitment manufacturers to retrieve,classify and manage the candidates,therefore,a skill label generation algorithm based on recruitment text information is proposed,the candidate labels are generated by extracting keywords through the recruitment lexicon,then the obtained text information is mined LDA the topic,and then the similarity between the topic keywords and the obtained candidate labels is calculated.Finally,the candidate labels are weighted and the skill labels under each topic are obtained.The experimental results show that the label generation mechanism can reflect the skill tendency of different positions and provide reference and basis for recruiting manufacturers to select candidates reasonably and efficiently.