Personalized Job Recommendation System Based On Mahout Framework

With the rapid development of information technology,the amount of network data increases exponentially.However,users are confused with overwhelming information loads and cannot quickly and accurately find what they really need in front of the massive data.The emergence of distributed computing,recommended systems and other related technologies effectively solve this problem.With the user preference model,the recommendation system can identify the users’ historical behavior data,and thus recommend valuable information to the users,which can reduce the time required for searching.This study describes the relevant theories about the recommendation system by introducing the Apache Mahout framework,Hadoop platform and other technologies,followed by summarizing and analyzing the existing recommendation algorithms.In addition,we illustrate the basic knowledge about non-additive measure and fuzzy integral in fuzzy theory and its application in the recommendation system.Furthermore,we propose a recommendation strategy model based on fuzzy integral.Finally,by integrating Hadoop distributed computing platform,combining Mahout and other open source technology framework,we design and implement a personalized job recommendation system based on Mahout framework and validate the system with simulation and test.The main work of this study is as follows:1.Propose a job recommendation strategy based on fuzzy integral.Briefly,based on the job-seekers’ attention to position indicators,we adopt fuzzy integral to calculate the attention degree of the position index by fuzzy comprehensive evaluation.Then we determine the job priority recommendation to the users based on the degree of conformity,which refers to the size of the comprehensive evaluation of the position.This study proposes and implements a job recommendation strategy model based on fuzzy integral.2.Design and implement a personalized job recommendation system based on Mahout framework,in which Hadoop platform has been used for distributed computing of large-scale job information and job recommendation system has been divided into user-related module,personalized recommendation module and distributed system module.