CareerMiner: Automatic extraction of professional network from large Chinese resume data

Abstract

The professional contacts of a person, including their past colleagues and supervisors, can play an important role in job recommendation and intelligent human resources management. However, collecting such information on a large scale can be challenging and costly. In this paper, we propose CareerMiner, a system to automatically construct a dynamic professional network to capture people’s colleague and supervisor relationships based on inter-related work experiences in a large collection of Chinese resumes. Specifically, we extract fine-grained work affiliation and job title information from resume text using a customized Name Entity Recognition model and introduce a trie-like data structure to efficiently index the hierarchical affinities of the implicit organizational structure in work experiences. Then, the topology of leaf nodes is used to identify professional relationships. Experimental results on real-world resume data are presented to demonstrate the effectiveness of the proposed system. Additionally, we develop an open-source demo system that includes the main features of the system. We also conduct a case study to illustrate how CareerMiner can be utilized for the social science research and data mining applications of professional networks.