Design And Implementation Of Micro-blog Sentiment Analysis Visualization System Based On Machine Learning

Microblogging,is a application platform which can be used to obtain information,Internet users through smart phones,tablets and instant messenger to share one hundred forty characters within a text message,and it has become more and more popular.this social way get through the gap between the Internet and mobile communications networks.With the mobile communication terminal explosive development today,people like microblogging and even rely on it.At the same time,large-scale micro-blog text presents new challenges and opportunities to the field of big data based on machine learning and natural language processing research.There are a large part of the text message which includes sentiment information in these text information.These emotional text messages are very valuable and important advice resources,how to use these micro-blog text which with emotional information to do the classification of sentiment,is a far-reaching,and challenging problem.And if we can design a scientific visualization display system,show the analysis result of the microblogging simply and intuitively,it will attract microblogging users,academic researchers,government policy-makers and business promotion to be used.In view of the above background,this paper studies the design and implementation of a visualization system of the analysis of microblog sentiment besed on machine learning,training support vector machines(SVM)classifier model,and ultimately realize the sentiment classification of all the text information,and using the visual microblogging information display system to show up.This paper introduced the theory of machine learning and micro-blog grasping technology,analyzed the business requirements and the functional requirements of the system,given the system application architecture design,using the Struts+Spring+Mybatis framework as project server,HTML5 provides simple and rich visual interface of the front operation,obtain microblogging data through Ajax asynchronous and converted the data instances into JAVA instance,converted to JSON format from the front display,use Mybatis persistence the framework in the same time.Using the MySQL database as the database of the system.The test results show that the system users can easily perceived microblogging users'attitude towards hot issues,obtain the dynamic microblogging sentiment real time,and the sentiment changes in different stages of attention to the event development,and it has the obvious practical significance in public opinion analysis of events and large-depth research to the support of big data.