Research On Consumer Intention Recognition Technology Based On Weibo Users

In recent years,the high-speed growth point has been concentrated in social media.The social media such as Weibo WeChat has sprung up,allowing people to share their lives,concerns about hotspots and current affairs.The popularity of social media has become more interconnected and shared.In this paper,Weibo is used as an example.Users express their ideas on Weibo through short texts of 140 words or less,and identify users' consumption intentions through deep mining and analysis of users' microblog texts.The definition of consumption intention in this paper is the purchase intention expressed by the user for the goods or services.By discovering the user's consumption intentions,the analysis and prediction results can be applied to various fields such as commercial promotion and performance prediction.For example,in the recommendation system,if the user is found to have a consumption intention,the user can accurately recommend the content corresponding to the user and create commercial value;when the movie is released or the product is released,the proportion of the consumption intention in the relevant text can reflect the user's Their needs provide the basis for the next judgment;when analyzing and predicting the user's consumption behavior,it is first necessary to determine whether the user has a consumption intention,and further research is valuable on the basis of judging the consumption intention.The traditional way of identifying consumer intent is to build a specific template to match according to different consumption intentions that the user may have.This paper collects the microblog texts of Weibo users in the field of film,and regards the consumer intent problem as a text classification problem.Firstly,the support vector machine is used to identify the consumption intention,which solves the complex and scalable implementation caused by building the template.Not a strong problem,and use the LDA theme model to extend the characteristics of the text.Then,deep learning is used to construct consumer intent recognition by constructing a model combining convolutional neural network and long-term and short-term memory network.In the process of consumer intention identification,the convolutional neural network extracts the features of the user's microblog text.As a result of the input of the longterm and short-term memory network,the long-term and short-term memory network processes the time series relationship of the features in the text,and outputs the classification of the consumption intention text.result.In the experimental results,the optimal F value of the support vector machine classification method is 70%,and the neural network model F value is 84%,which can effectively identify the user's consumption intention.Then introduced the affair atlas technology,through the analysis and prediction of events,constructing an instance map to reveal the evolution of things,to better mine the intention of consumption and to accurately push products or services.