Research And Realization On Personalized Post Recommender System

In online system of talent recruiting, because of the information asymmetry between the job seeker and recruiters, the former spends lots of time and energy to select a suitable post, so a dilemma situation of difficulty to find a post is formed. Traditional way of post recommendation can't satisfy the need of current recruiting system, so a new way based on personalization emerges. The post recommender system based on personalization can conduct fuzzy intelligent match according to the job seeker's ability structure, professional quality, application intention and so on to provide satisfying services for users.Based on the research of the key technology of the current personalized information service field, the paper proposed a personalized post recommender system based on Multi-Agent. The new system uses roulette model to characterize the users' job seeking intention, so among each characteristic item has a richer semantic meaning, in addition, the system apply the improved algorithm based on the domain knowledge combination of content and collaborative filter recommendation, finally, the recommendation result is more accurate. The Multi-Agent technology improves the system's intelligence and robustness, ontology enhances the system's availability and expansibility.At last, the paper conducts a comprehensive performance evaluation, further confirms that the proposed system has a obvious advantage in technology and a greater value of application and extension.