Research On Collaborative Filtering Recommendation Service Based On Social Media Gene Map

The rapid development of information technology has laid the foundation for the rapid rise of social media.Users have multiple identities.Social media mainly serve users.According to the categories of social media,appropriate recommendation methods are adopted to provide users with personalized services.However,most of the existing recommendation methods take the similarity between users or projects as the dimension to make user recommendation or project recommendation,and the calculation method of similarity is relatively single,which cannot adapt to the recommendation research of large-scale network data.Or in the research methods,they either lay more emphasis on interests or content labels,rarely taking into account a variety of influencing factors at the same time,and the accuracy of recommendation needs to be improved.To solve the above problems,this thesis proposes a improved collaborative filtering.The following are the main research contents of this paper.First of all,based on the related theory in biology,through social media user behavior data,this paper specifications the biological gene,social,and gene network,calculate the core gene compatibility on genes,using the entropy weight method to determine the weights of genes,genome construction of social media,and implement recommendations.Then,in view of the current situation that social media recommendation methods mainly focus on user or item recommendation,this paper introduces social media gene map and project similarity into weighted Slope One algorithm,and proposes a collaborative filtering recommendation algorithm based on social media gene map.In the measurement of item similarity,a method combining score similarity and semantic similarity is adopted.It is introduced into the weighted Slope One algorithm together with the user's genetic matching degree to realize the comprehensive service recommendation.Finally,for the optimized algorithm,sina weibo is taken as the data set,and the two algorithms before optimization and the other two optimized algorithms are compared and tested respectively.According to the data of MAE and MRSE,the recommended algorithm in this paper has been significantly improved.