Multi-granularity and variable weight evaluation model for person-job matching

It is a prerequisite to conduct the person-job matching evaluation for human resource management. The existing matching evaluation models mostly adopt the fixed weight method, and ignore the differences in the personnel structure of different units, and cannot adapt to the coarse and fine changes of the resolution granularity caused by compiled data. Therefore, a multi-granularity variable weighting evaluation model is constructed. Firstly, for the ambiguous data of resource allocation and person-job matching, the comprehensive index method is applied to extract four main factors: job position, grade, expertise, and quantity. A unit’s technological level evaluation index is proposed, classifying units into three categories: specialized efficiency-oriented, quantity scale-oriented, and hybrid models. By applying the coefficient of variation-G1 method, a variable weighting expression for the matching degree index of different unit types is formed, thus constructing a coarse-grained variable weighting evaluation model. Secondly, for fine-grained resource allocation and person-job data, considering the differences in grade, expertise, job position, and other aspects, the impact of different factors such as grade, expertise, and job position on weights is quantified. The concept of "matching deviation" is proposed and defined to comprehensively evaluate person-job matching. It leads to the construction of a fine-grained evaluation model. Finally, the evaluation under different resolution granularities is realized according to the real data. It verifies the effectiveness of the method model.