Design And Implementation Of User's Feature Analysis Platform For Social Media

In recent years,relying on the rapid development of the Internet,social media emerges unexpectedly.The traces left by users during the use of social media imply great value.At present,relevant researches mainly focus on user's basic attributes and behavioral attributes,while fewer researches on the personality or the social attitude have been done.The analysis on the personality characteristics is of great significance for online public opinion management,key population monitoring and other activities.However,the user's personality information is expressed invisibly by the usage habits of words.Therefore,how to establish the mapping relationship between the user's text data and the personality characteristic is the main difficulty in relevant researches.At present,the research on related directions is mainly based on regression algorithms,while the social media targeted and the algorithm model used by researchers are different.In the thesis,the gray relational degree analysis is used instead of correlation analysis for feature dimension reduction,so as to deal with the nonlinear scene better.And BP(Back Propagation,BP)neural network is applied to the study of personality analysis,which makes full use of the mapping ability of multi-dimensional function of neural network,and completes the work of analyzing user personality through user text.At the same time,the thesis designs and builds a feacture analysis platform for social media,completes a set of interactive WEB system consisting of front-end,backend and database.The system can analyze the personality of weibo users based on their weibo text and display the analysis results visually.The work in the thesis mainly focuses on the following aspects.First,the paper clarifies the design intent and business requirements of the platfonn through requirement analysis,and on this basis,carries out the technical selection.Second,the BP neural network model was trained and implemented to predict user personality.In the data preparation part,the user's personality information and the user's text data are collected.Then,guided by the characteristic of the data,BP neural network is used to implement the algorithm.In this process,feature engineering is established to filter features for modeling,and appropriate parameters for modeling are also gradually explored.Finally,the satisfactory training result is obtained.Third,according to the specific needs,we complete the design and construction of the platform and realize the functions.In the work,we first fix the general idea,then complete the realization of each function according to their requirements,and finally we test the platform.In the end,the thesis summarizes the deficiencies in current work and points out the direction for further transformation and impro-vement.