• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

计算机工程与科学

• 人工智能与数据挖掘 • 上一篇    下一篇

基于机器学习的能力评价与匹配研究

张毅,张珉浩   

  1. (重庆邮电大学通信与信息工程学院,重庆 400065)
  • 收稿日期:2017-12-25 修回日期:2018-01-11 出版日期:2019-02-25 发布日期:2019-02-25

Competence evaluation and
matching based on machine learning

ZHANG Yi,ZHANG Minhao   

  1. (College of Communication and Information Engineering,
    Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
  • Received:2017-12-25 Revised:2018-01-11 Online:2019-02-25 Published:2019-02-25

摘要:

目前,高校学生就业形势严峻,针对企业看重的能力对学生做出评价,有助于企业选拔人才,同时也能提高学生的竞争力。
采用层次分析法和模糊评价相结合的方式对高校学生的综合能力进行评价。为了解决模糊系统无法自动调整隶属函数参数的问题,有效结合模糊理论和神经网络架构的优点,提出了引入神经网络的综合评价改进算法;并设计具有时频局域化特性的小波神经网络,能够更好地模拟非线性函数,用于预测学生适合的职位。分析实验结果表明,基于改进模糊神经网络算法的能力评价模型与小波网络职位匹配模型,能够提升系统精度与自适应能力,评价结果客观,对学生的能力评价及就业选择具有指导意义。
 

关键词: 机器学习, 综合能力, 模糊评价, BP神经网络, 小波神经网络

Abstract:

At present, the employment pressure of college students is severe, so there is an urgent need to evaluate the capabilities that enterprises value. This is conducive for enhancing the competitiveness of students and helps employers select qualified personnel. Firstly, we combine the analytic hierarchy process (AHP) and fuzzy evaluation to evaluate college students' comprehensive quality. In order to solve the problem that the fuzzy system cannot automatically adjust the parameters of membership functions, effectively combining the advantages of the fuzzy theory and neural network architecture, we propose an improved comprehensive evaluation algorithm. We also design a wavelet neural network with time-frequency localization to better simulate nonlinear functions and predict suitable positions for students. Experimental results show that the competency evaluation model based on the improved fuzzy neural network algorithm and the wavelet network job matching model are able to improve system accuracy and self-adaptive ability, and the evaluation is objective, which have a guiding significance for the competence evaluation and employment options of the students.
 

Key words: machine learning(ML), comprehensive quality, fuzzy evaluation, BP neutral network(BPNN), wavelet neutral network(WNN)