基于协同过滤算法的人力资源信息管理系统研究  

Human resource information management system based on collaborative filtering algorithm

作　　者：张玙[1] ZHANG Yu(Business School Hohai University ,Nanjing 211100, China)

机构地区：[1]河海大学商学院,江苏南京211100

出　　处：《电子设计工程》2017年第3期23-27,共5页Electronic Design Engineering

基　　金：国家自然科学基金资助项目(61303005);国家社会科学基金资助项目(10bzx73)

摘　　要：为了方便企业及时便捷的获得与追踪各种人事管理方面的信息,在Hadoop分布式构架平台的基础上,利用员工对企业的项目满意度构建行为记录矩阵,通过协同过滤算法分析和设计了一套适用于人力资源信息管理的系统,实现参数设定、分区管理、职员管理、用户管理和生产管理五大模块的综合管理。对人力资源信息管理系统进行了具体的实验分析,结果表明:当1 571名员工产生28 325条员工评分记录时,2个节点和3个节点的系统响应时间分别为7.4 s和4.8 s,且明显小于非分布式算法,同时协同过滤算法比非分布式算法的MSE与MPE分别小0.39%、0.52%,本研究的协同过滤算法可以胜任人力资源管理系统的数据分布式计算任务。In order to facilitate timely and convenient access to business and tracking a variety of personnel management information, based on the Hadoop distributed architecture platform, the use of staff of the enterprise satisfaction with the program to build a matrix acts recorded by the collaborative filtering algorithm analysis and design a set of It applies to human resources information management system, parameter setting, partition management, staff management, integrated management of user management and production management of the five modules. Human resources information management system specific experiments, the results showed that： when 1 571 employees produced 28,325 employees score recorded two nodes and three-node system response time was 7.4 s and 4.8 s, and significantly less than non-distributed algorithms, collaborative filtering algorithm, respectively, while smaller than the non-distributed algorithm MSE and MPE 0.39%, 0.52%, this research collaborative filtering algorithm capable human resources data management system for distributed computing tasks.

关 键 词：协同过滤算法 Hadoop分布式 信息管理 系统设计 人力资源 

分 类 号：TN18[电子电信—物理电子学]