A Graph Theoretic Approach for Employee Performance Appraisal

In an organization, employees are often evaluated (ranked) by multiple supervisors (or co-workers) in different time periods. Providing a single ranking based on these evaluations in such a way that it has the most compatibility and consistency with the provided evaluations is called rank aggregation. This rank aggregation problem, which has wide applications in various sciences, has been studied extensively theoretically in mathematics and computer science. In this paper, we study for the first time, the application of rank aggregation in the field of performance appraisal. In this application, the performance rankings of employees are given by a number of evaluators and the goal is to provide a final ranking for the employees that has the most similarity to the given evaluations. Other than obtaining the most consistent rankings, we show that more inference results like the quality of the evaluators and informal relationships between individuals can be obtained as by-product of this process of rank aggregation.