Research And Application Of Recommendation Technologies Based On Agent

As the process of information technology and the increasing popularity of the internet,people can enjoy convenience services through the intemet without leaving home.At the same time,people assume the pressure of "information overload".It causes for wide attention that how to provide rapidly and exactly services for each user to meet the actual needs of the personal information.Recommended technology provides a better solution to solve this problem,it forecast user's interest through recommendation algorithm,recommend the resources which the users like,filter out what users don't like.It can enable the web's goods,news, advertising,etc to the greatest degree of promotion to a specific user.Recommended technology has changed the way that people access to information.From active to passive acceptance of retrieval,the user can quickly and efficiently find favorite resources through recommendation of the website,and ultimately improve the user's satisfaction and loyalty.People's interest changes over time,and people's memory forgets gradually over time.Traditional recommended technology can not reflect that time have an impact of changes in user's interest and the effect of forgettingof memory.This paper is based on the traditional collaborative filtering algorithms and Ebbinghaus forgetting memory theory,analyse the user's access sequence with the help of time-weighted method,and recommend through linear forgetting function and normal forgetting function.In this paper,the improved recommendation algorithm can not only reflect the affect of time to changes of user's interest,but also recommend the information regularly according to the degree of forgetting memory,in order to deepen user's memory.Agent can automatically sensing the changes in the surrounding environment, response autonomy and affect the surrounding environment.In this paper,agent technology will be recommended in the information field,realize intelligent automation of data collection,analysis,processing and recommendation.In the environment of network node,agent can carry the necessary code to run,find a free host and move to it for recommend computing.This will effectively save network bandwidth and make full use of network resources of the host.Finally,this paper puts forward the architecture based on the recommendation of Agent system,design principles and work flow,realize the recommendation system and test it.