数据表示和特征修正在人岗匹配研究中的应用

Application of data representation and feature modification in resume-post matching research

人岗匹配度的合理测算是人才合理分配的基础,针对传统人岗匹配研究中主观评价占比高、数据表示粗糙等问题,对人岗匹配度测算方法进行了研究。首先构建了人岗匹配度指标体系;其次进行特征构建得到特征字段属性,采用数据表示和特征修正的技术方法进行特征表示、多区间划分及修正,完成各指标匹配度的测算;最后采用XGBoost算法构造人岗匹配度分类模型,将模型概率值作为匹配度测算值,通过真实招聘数据对该方法的有效性进行了验证。

The reasonable estimation of resume-post matching degree is the basis of rational allocation of talents.Aiming at the problems of high proportion of subjective evaluation and rough data representation in the traditional resume-post matching research,the estimation method of resume-post matching degree is studied.Firstly,the index system of resume-post matching degree is constructed.Secondly,the feature field attributes are obtained by feature construction;the technical methods of data representation and feature modification are used for feature representation,division,modification,and the calculation of the matching degree of each index.Finally,the XGBoost algorithm is used to construct a classification model of resume-post matching degree and the model probability value is used as the matching degree estimation value.The effectiveness of the method is verified by real recruitment data.