Design And Implementation Of Film Recommendation System Based On Deep Interest Evolution Network

As people's lifestyle becomes more and more rich,large-scale users also produce a larger scale of application data.To solve the problem of information overload,recommendation engines have been created.However,at present,most recommendation algorithms directly regard the interest behavior as the result of interest,while the potential interest is difficult to be fully reflected through the explicit behavior.Most methods ignore the real user interests behind the mining behavior,and the user interests are constantly changing,so it is very important to obtain the dynamic relationship between the interests for the expression of interest.Therefore,further research is needed to solve these problems.In order to better capture the dynamic relationship between users' interests and accurately predict the change process of users' interests,the system proposed to use Deep Interest Evolution Network to capture the dynamic relationship between interests and predict the Evolution process of users' interests.The deep interest evolution network is mainly divided into three layers.The first layer is the behavior sequence layer,which converts the original data into the Embedding behavior sequence.The second layer is the interest extraction layer.The specific work of this layer is to simulate the evolution process of users' potential interest,and then extract the sequence information of users' interest features.The third layer is the interest evolution layer.On the basis of the interest extraction layer,the self-attention mechanism is introduced to better simulate the interest evolution process related to the current movie target.By comparing the improved algorithm with the original algorithm,it is found that the improved algorithm can better capture the dynamic relationship between interests.Finally,several other mainstream recommendation algorithms are experimented,and the improved algorithm is found to be more accurate than other algorithms through comparative analysis.Finally,based on the proposed algorithm,this paper designs and implements an online recommendation system,which uses Spark platform to process data,then uses Redis to save the processed characteristic data,and finally deploys the system on Jetty.After testing the system module,the usability of the system is verified.