Construction And Research Of University Students Personalized Employment Recommendation System Based On Hadoop Platform

In the 21 st century, we take the high-speed trains named Internet technology, and enjoy great convenience and bonus that the Internet brings. At the same time, large scale data generated by the Internet has caused information overload problem. Using search engine, we can retrieve the relevant content, but can’t meet the demand of our personalized data; The personalized recommendation system enters into our field of vision widely and has been become a powerful tool to solve the problem of information overload. However, based on stand-alone the personalized recommendation system still can’t provide efficient and fast personalized service in the face of vast amounts of data, so you need to combine the Hadoop platform for dealing with large-scale data effectively with personalized recommendation system to provide satisfying personalized recommendation service under the environment of huge amounts of data. Through investigation and analysis, a large number of employment recruitment information was published by a third party web site or employment information was included among them; In the difficult employment environment and plenty of employment information around them, because fresh graduates own relatively narrow channels of employment and their personal career pursuit is not very clear, they feel confused and have difficulty with looking for a decent work which is according with the actual situation of individual. On the above issues, it is very necessary to construct a university students personalized employment recommendation system based on Hadoop platform.On the basis of predecessors’ research, according to the actual conditions of employment information recommendation, this paper completes the following work:1)Judging by the actual needs, a personalized employment information recommendation system that can handle the massive data has been designed, including personalized recommendation engine module, user related module and distributed system module. Combined recommend online with offline data analysis and calculation, It can provide employment information for university students in line with the actual situation of their personal effectively and quickly.2)According to the characteristics of MapReduce and Mahout and theoretical research of personalized recommendation, combining the practical application environment, content-based recommendation was improved. Make it not depend on the historical data, which can overcome the recommendation system cold start problem.Then the improved content-based recommendation and items-based collaborative filtering recommendation constitute parallelized hybridization module to overcome the data sparseness of recommendation system.3)Confirmed by simulation test and system test, this system can solve the problem of employment information overload, can take "people looking for work" mode into “job matching people”, can provide personalized employment information recommendation service for university students.