Research On The Technology Of Network Hotspot Foundation Based On Micro-blogging

With the rapid development of Web technology, Source of content in the form ofmicro-content on the Internet has gradually established a significant advantage. TheMicro-Blog as a form of media of the Internet, with its short, convenient features showexplosive growth dynamics, Since the Micro-Blog writing low threshold, plus releaseconvenient to instantly share information dissemination time tends to zero, has become thehot events generate, disseminate important source, The Twitter influence also presents thegeometric multiplier trend and penetrated into all aspects of society at an alarming rate.Micro-Blog has become the second largest source of public opinion, and to play animportant role in the transfer and diffusion of public opinion, and emergencies.Micro-Blog simple text (usually no more than140words) updated information and avariety of tools to publish and share information Micro-Blog showing fragmentation-in-time,mobile and other characteristics, and is no longer intact information content, Micro contentplus internet has wide sources, update, participatory and interactive features, some of theexcesses of the conversation is easily transmitted even blind manipulation or use of, if notactively control and response, negative emotions will snowball gradually become larger, thegovernment, businesses or other institutions into the cusp, Therefore Research, discovery,monitoring and management of Micro-Blog hot events all the more important.1.This article is mainly generated from the micro-content background and significance,related research status begin on the urgency and inevitability of this research, monitoringpublic opinion analysis technology, designed based on a short text clustering model, and adetailed description of the concepts and technologies related to this model.2.In this paper, the characteristics of short text design the Micro-Blog information flowsession decimation algorithm to overcome the short information incompleteness andstaggered; TF-IDF method in the similarity measure with good results but when the text isrelatively short, and will reduce the number of matches between the word in the text, so thatthe similarity drift, for this improved methods of the TF-IDF, to overcome the Keywordssparsity similarity drift; Design of short text clustering algorithm can effectively mixedMicro-Blog text clustering and can meet the requirements of accuracy and scalability.3.Finally, public opinion monitoring system based on a short text on the design ofexperiments, including: session decimation algorithm and improved TF-IDF similarity measure method is applied to the design of mixed clustering algorithm analysis, and basedon the experimental results of the performance evaluated experimentally verify thefeasibility of the method and experimental results.