The Research On Topic Access And Evolution With LDA

With continuing advances in Internet technologies for information dissemination means,Internet users can more easily get from the network within a certain area of a large number oftext messages. But access is not the same as knowledge, in the endless world of information,subject to the limited time and space, anyone can feel helpless. Therefore, complex textquickly and accurately obtain information we crave knowledge and information, becomes animportant issue that must be solved in the current situation. For the information in the text,separately from the perspective of two different macro and micro analysis.We can assume that a text is composed of several subjects, these topics reflect the mainelements of a text, how can I obtain these topics and learn about topics changes over time, iswhat issues need to be addressed in this study. Therefore, this study aims to use moderninformation technology at a macro-level automatic analysis on the topic of evolution, madesubject acquisition and evolution of the LDA model based methods, secondary topic inpeople’s grasp and understand large amounts of text information, and related topics for thedevelopment trend of evolution of technology topics.Evolution is a reflection of a topic from the topic, to become a hot spot, and graduallyfaded out, including changes in intensity and content. How to automatically generalize andanalyze large amounts of text in a hidden topic of evolution is the focus of many researchersare concerned. Topics of evolution is the process of dynamic development on the topic at thetime. On topic already exists for this document in the acquisition and evolution oftechnologies, mainly on the basis of meaningful exploration of the following:(1) Propose a method to dynamically select topic number, you can determine the numberof different topics in different time slices, main subject can lead to excellent quality, the topiceasier to understand.(2) Frivolous conversation on the topic of evolution has a negative impact, it will affectnot only topic generating quality and accuracy for topic evolution or even incorrectly guidedevolution of the topic. This article focuses on insignificant topics filtering, filtering outinsignificant topic, whether the topic as a whole is more readable, and authenticity is the topicof evolution could have a positive role in promoting.(3) Finally also tried to get this article topic topics quality, intensity and contentevolution automatic evaluation of the results. Journal of Chinese information and the information of the journal experimental resultsshow that our topic extraction based on topics filtering topic and evolution of methods can beobtained more accurately on the strength and content evolution.