Research On Personalized Recommendation Method Based On User Psychological Mining And Analysis

With the rapid development of the Internet,the number of information shows exponential growth.People accept the overloaded information actively or passively in their daily life and work,which invisibly increases the cost of their life and work,therefore,an invisible contradiction comes into being between the value of unit time and search value of unit time.Based on the above problems,in order to adapt to the rapid development of the Internet,the recommendation system emerges as the times require,the main role of the recommendation system is to achieve intelligent recommendations.The traditional recommendation systems improve themselves by studying recommendation algorithms,regarding the user resource information as static,immutable resource for the operation among the users' projects,which improves the performance of the recommendation system,but can't greatly meet people's psychological needs.With the continuous development of the Internet,the computer technology makes great progress,the personalized recommendation method which is based on users has become an important research object of scholars,and has realized the transformation from simple algorithm optimization to the combination of algorithm and user psychology.In real life,people's psychological activities affect the recommended effects of the recommendation system,because people's psychological activities will be affected by their own needs and external factors,there is a state of psychological instability,in order to achieve a true personalized recommendation,mining and analyzing which are based on the user psychological become particularly important.This article carries on the psychological excavation from the users' behavior,forgetting and the interest characteristics,through domestic and foreign research and analysis on users' psychological activities,the users' dynamic psychological activity is the important factor that affects the system recommendation.Besides,this article also puts forward related ideas which are relevant to the users' psychological analysis,The common method is to quantify the factors that affect users' psychology,predicting Users' psychological trends dynamically,making accurate recommendation by combining relevant recommendation methods.Firstly,this paper makes an intuitive behavior information extraction from the users' behavior characteristics,and puts forward a group behavior grouping method which is based on RFM model.This method can classify users by explicit behavior,excavate through incremental users' behavior habits,avoiding the problem of classification difficulty under big data to some certain extent,and realize the automatic acquisition of user behavior data.Secondly,from the research of the inherent law and dynamic interest of the users' psychology,this paper puts forward a dynamic interest model based on the users forgetting curve,makes a quantitative analysis from the user's forgetting enhancement aspect,predicts the users' interest degree through the change of interest,and finally combines the dynamic interest model based on the user group behavior and the forgetting curve.The clustering algorithm realizes the user aggregation and the collaborative filtering type recommendation.Through the correlation comparison experiment of Movielens dataset,it concludes that this method has strong applicability and recommended value.