Exploring the optimum pattern for knowledge workers selection using DEA and CART compilation approach

The success or failure of any organization is directly linked to the quality of its human resources selection (recruitment, measurement, and selection). By reviewing the data of a knowledge job, this paper aims to help improve the selection process of that job. Consequently, the selection of appropriate employees’ rate will increase, and the rate of human resource turnover will decrease. The approach of this paper is Applied Research and the strategy is Case Study. This paper combines two computational techniques (DEA and CART). Data Envelopment Analysis (DEA) is a non-parametric technique that determines the efficiency of individuals, but it does not provide information on the details of factors affecting performance (especially non-numerical factors). In the present study, this deficiency has been resolved using the Classification and Regression Tree (CART) (as a data mining technique). The result of this study has provided a framework for combining DEA and CART in order to discover rules on the recruitment of knowledge workers in a specific job (a knowledge job) and in a specific organization (HFJ Institute). The results indicate that ‘work experience’, ‘average score in the last degree’ and ‘age’ are related to the employee performance, and therefore it is necessary to be considered in the process of future recruitment of that job.
By reviewing the data of a knowledge job, this paper aims to help improve the selection process of that job. Consequently, the selection of appropriate employees’ rate will increase, and the rate of human resource turnover will decrease. The approach of this paper is Applied Research and the strategy is Case Study.
In the literature review section, the definition of recruitment and selection (Azar et al. 2013), the definition of knowledge workers (Drucker 1994; Horwitz et al. 2006; Li et al. 2015), tasks of human resource management (Osman et al. 2011) and a background of the use of data mining in the field of human resources (Hajiheydari et al. 2017) have been reviewed.
This paper combines two computational techniques (DEA and CART). Data Envelopment Analysis (DEA) is a non-parametric technique that determines the efficiency of individuals, but it does not provide information on the details of factors affecting performance (especially non-numerical factors). In the present study, this deficiency has been resolved using the Classification and Regression Tree (CART) (as a data mining technique).
  In this research, we tried to develop the previous models and present a new model.
The result of this study has provided a framework for combining DEA and CART in order to discover rules on the recruitment of knowledge workers in a specific job (knowledge job) and in a specific organization (Hedayat-e Farhighteghgan-e Javan (HFJ) Institute). The combination of data envelopment analysis and data mining approaches (and considering qualitative and implicit variables in the estimation of efficiency) is one of the most important innovations in this research.
In the proposed framework, organizations can identify and recruit talents and appropriate individuals in a short time based on data mining and discovery of success patterns (resulted from their past experiences). This action avoids costs of frequent recruitment, and decreases turnover rate and improves performance.
By analyzing the outputs of the designed model for the stage of recruitment and selection in this specific job (a knowledge job) and specific organization (HFJ Institute), six rules were extracted, and based on that, suggestions were given.
At the stage of recruitment, it is better for this organization to take into consideration these rules (the status of jobseekers in ‘work experience’, ‘average score in the last degree’, and ‘age’) and then to decrease costs and failure rate of the recruitment process.
The results are somewhat consistent with the results of previous studies. The proposed approach can be planned and implemented in various jobs and organizations to extract specific rules for these jobs and organizations in order to increase productivity in the process of human resource selection and recruitment.
References
Zhu, X., Seaver, W., Sawhney, R., Ji, S., Holt, B., Sanil, G. B., & Upreti, G. (2017). Employee turnover forecasting for human resource management based on time series analysis. Journal of Applied Statistics, 44 (8), 1421-1440.
Lukovac, V., Pamučar, D., Popović, M., & Đorović, B. (2017). Portfolio model for analyzing human resources: An approach based on neuro-fuzzy modeling and the simulated annealing algorithm. Expert Systems with Applications, 90, 318-331.
Osman, I. H., Berbary, L. N., Sidani, Y., Al-Ayoubi, B., & Emrouznejad, A. (2011). Data envelopment analysis model for the appraisal and relative performance evaluation of nurses at an intensive care unit. Journal of medical systems, 35 (5), 1039-1062.