系统管理学报

系统管理学报 ›› 2015, Vol. 24 ›› Issue (3): 365-371.DOI: F 272.3

• 管理信息系统 • 上一篇    下一篇

基于文本挖掘的众包人才能力分析

刘景方1，张朋柱2，吕英杰3, 张晨2   

  1. 1.上海大学 管理学院，上海 200444；2. 上海交通大学 安泰经济管理学院，上海 200052；
    3.北京化工大学 经济管理学院，北京 100029
  • 收稿日期:2014-03-03 修回日期:2014-06-06
  • 作者简介: 刘景方( 1980-) ，男，讲师，研究方向为信息管理。E-mail: jingfangliu@shu.edu.cn
  • 基金资助:

    国家自然科学基金资助项目（71301102，71171131）

Analysis of the Competence of Crowdsourcing Talents using Text Mining

LIU Jing-fang1, ZHANG Peng-zhu2, LU Ying-jie3, ZAHNG Chen2   

  1. 1. School of Management, Shanghai University, Shanghai 200444, China; 2.Management Information System center, Shanghai JiaoTong University, Shanghai 200052, China; 3. School of Economics and Management, Beijing University of Chemical Technology, Beijing100029, China
  • Received:2014-03-03 Revised:2014-06-06

摘要： 由于众包模式中存在人才盲目参与任务竞争的问题，迫切需要分析清楚众包环境下人才所需要具备的能力。同时，众包平台上出现很多人才之间进行经验交流的网上社区，面对海量的人才众包经验数据，急需通过有效的方法来获取众包人才能力。针对网上众包社区中的经验沟通交流信息，通过文本挖掘技术来分析众包人才的能力。为了从众包人才交流社区的非结构化文本中识别出与能力特征，基于文本聚类的主题识别方法，将人才交流文本内的句子按其不同主题进行聚类，每一个结果簇表示某一种能力特征，采用基于关键词的聚类结果表示方法来解析每一个簇。通过对聚类结果的分析，分别确定了众包人才能力的5个方面：学习与创新能力、服务意识、在线社交能力、成就导向和竞争意识。通过实验检验了所提出方法的有效性。最后，分别分析了程序开发人才和标志设计人才的能力差异化原因。

关键词: 文本挖掘, 众包, 人才, 能力

Abstract: Since the talent is involved in blind task competition in crowdsourcing, it is necessary to analyze the competence of talents in crowdsourcing. Meanwhile, there are a lot of online communities where talents can exchange experience. With the massive empirical data, an effective way is urgently needed to obtain the talent competence. Using the experience communication data in an online crowdsourcing community, we adopt the text mining technology to analyze the talent competence in crowdsourcing. To identify the competence from unstructured text in crowdsourcing talents communication communities, we use topic identification method based on text clustering. The sentences in talent exchange text are clustered as different themes, each cluster representing a competence. A keyword-based clustering result representation method is used to analyze each cluster. By analyzing the clustering results, we identify five crowdsourcing talent competences: learning and innovation, service awareness, online social skills, achievement -oriented and competitive spirit. The effectiveness of the method is evaluated by experiments. We finally analyze the reasons of differentiation of crowdsourcing talent competency.

Key words: text mining, crowdsourcing, talent, competence