Research On The Influence Of Social Network Users In Stock Fields And System Realization

With the Internet and social media and other technology to develop,people are willing to spend more time online communication and interaction.Social media has become one of the most important platforms in the exchange of information,as is the case in the financial sector.Ordinary investors and professionals engaged in securities investment will make some comments and information on social media.However,social media between people will interact with each other,especially in the securities sector,the impact of high people will affect the less influential people on the securities market views.We know that there is a close relationship between "public sentiment" and "market sentiment",so it is imperative to find influential people in the field of securities.The traditional research method is mainly based on the network topology,based on user behavior or content-based to build user influence model,but this model is single,unstable,easy to lose important information.This article mainly takes the financial social platform as an example,according to the user's comment and the related behavior data,carries on the influence analysis research.Most of the commentary data contains the views of the securities investors on the market,through the introduction of natural language processing technology and depth of learning methods for mining,and with the actual securities market comparison,to build the main characteristics of the model.In addition,this paper also considers the characteristics of user activity and user behavior,and proposes a new user influence model-UASRank model.The content and innovation of this paper are as follows:(1)The LSTM-based emotion analysis model is proposed,and the natural language processing technology is introduced to carry out the feature optimization and model improvement.The experimental results are compared with the emotion analysis model based on the emotion dictionary and the emotion analysis model based on SVM.The experimental results show that the LSTM-based emotional analysis model has the highest accuracy.(2)Proposed a new user influence algorithm-UASRank algorithm.UASRank algorithm based on the PageRank algorithm focuses on three parameters:1)user activity,that is,the user in a period of time posting frequency;2)user interaction,that is,the user interaction with the frequency of fans;3)The accuracy of the point of view,that is,the user view and the market market matching degree.The first two parameters can be obtained through statistics,and the user view of the accuracy of the views of the user published by the above mentioned based on the learning model of emotional analysis of emotional tendencies,and then emotional tendencies and market comparison,get the user view accuracy.(3)To build a UASRank algorithm based on the user influence ranking prototype system to provide more investment experience and stronger market pre-judgment ability of successful investors to the general user attention to help ordinary investors to obtain more valuable investment advice and information.