Implementation Of User Portrait Based On Tag Propagation

With the rapid development and increasing popularity of the computing Internet,users have put forward higher requirements for the construction of portal websites.However,not all portals' content design meets the user's requirements for the website,resulting in a lot of inconvenience to the user during use,the user experience becomes worse,and the user eventually loses.If you want users to be more sticky to the website,you have to do what you want,according to the user's usual relevant browsing data feedback,understand the user's behavior habits,optimize the website in a targeted manner,and accurately grasp the users.With the rapid development and increasing popularity of the computing Internet,users have put forward higher requirements for the construction of portal websites.However,not all portals' content design meets the user's requirements for the website,resulting in a lot of inconvenience to the user during use,the user experience becomes worse,and the user eventually loses.If you want users to be more stick to the website,you have to do what you want,according to the user's usual relevant browsing data feedback,understand the user's behavior habits,optimize the website in a targeted manner,and accurately grasp the users.Based on the related browsing data obtained by a wifi portal user,this paper proposes a method for constructing a user portrait based on the existing unreasonable design of the website.Through a series of processing,the user's web browsing behavior data Modeling browsing behaviors,constructing user tags,and displaying user portrait results,website managers can use these display results to get some specific suggestions,which can solve some problems that are not fully considered when the website is designed,making the website design more Reasonable and more practical.The user portrait system based on label propagation uses the point density-based label propagation algorithm(NDLP)to analyze user behavior and cluster users.With the help of the clustering center,each type of user can be abstracted into refined short text.The user is a collection of short texts with highly refined semantics.The analyst can quickly recognize the user through the user portrait,understand the user's personal preferences,and also promote the user's secondary classification.Behind every refined short label is the analysis and mining based on a large amount of data,so theusability of the data is improved to a certain extent,so that the analyst is free from facing a large number of complex and irregular raw data.This user portrait system is divided into three layers: data layer,processing layer and presentation layer.After the data in the data layer is acquired,a simple preprocess is performed.The processing layer analyzes the data in the data layer to obtain the results of the user portrait.The presentation layer presents the result of the processing layer in different forms for the data analyst to perform specific results analysis.