Design of employer branding analysis tools based on network mining algorithms

• Main Authors: Lee, Kuei-Lin, 李魁林
• Other Authors: 沈錳坤
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/cy332b

碩士 === 國立政治大學 === 資訊科學系碩士在職專班 === 106 === Employer Branding is about how attractive an employer is to its talent pool and assures that current employees are engaged. The current development of the measurement methods of employer branding are qualitative and quantitative research methods. The limitation of these approaches is the coverage of companies and costs of studying. This study presents a set of tools for employer branding analysis from large-scale resume database based on Backhaus & Tikoo’s Employer branding framework. The proposed approach constructs the Enterprise Endorsement Network which is a weighted directed graph of enterprises and the edge is generated from career transition records. Then network mining algorithms are employed to assess the employer attractiveness of enterprises. Furthermore, for employer brand loyalty, a probabilistic employee retention model is proposed and similar retention models of enterprises are clustered to build the baseline models for comparisons of retention tendency. Experiments on resume database for enterprises of Taiwan show the usefulness of the designed employer branding analysis tools.