Research On The Application Of Natural Language In The Field Of Military Public Opinion

In recent years,internet has penetrated into every aspect of our daily life,every intelligent terminal can be a carrier of information,the mobile Internet has become an important tool for the daily life of people.In the environment where the Internet is becoming more and more popular,people's freedom of speech is getting higher and higher,peoples can browse the news or make comments anytime and anywhere,BBS has become a new place for people to speak.These platforms carry huge amounts of information.In the field of demilitarized-war,information-war is a new form of war with no gunpowder smoke,network platform is becoming the new battlefield of information warfare,public opinion on the Internet is related to the political security of the party and the country,network-public-opinion-war which resulting from public opinion has become the commanding point of great power game,it is an important factor to win the information war.In this age of information explosion,the truth is often drowned out in a number of induced emotional messages that are intended to be known to the Internet,the cost of verification and verification of information is far higher than the cost of information editing,it is the reason that in the public opinion wars of all countries,the attacking party is often on the initiative,and the defenders are tired of dealing with it.The complicated public opinion shows that internet is the main battleground of public opinion,it is likely to be the "biggest variable" facing contemporary China,even "the heart disease".At present,the major military BBS has become the main place for the people to publish their military statements,the articles and comments in each section contain a large amount of military public opinion information.So,how to make good use of these speech information,grasp the public opinion direction,and turn the passive situation of the public opinion war in our country,has become the problem which need to face and solve in current.Using data mining technology to analyze text content is a hot topic in the field of big data research,the natural language processing technology evolved from this has been widely used in public opinion analysis of WEIBO and BBS,the major websites can monitor the public opinion information in real time and grasp the sensitive topics and the emotional inclination of peoples in a timely manner.This paper takes the topic and relevant comment data as the research object from the public BBS of super camp,design and implement the public opinion analysis system in the field of military public opinion,the main research work and research results as follows(1)Data collection.Using spider technology to climb up posts and related review data for army and force and navy and police and aerospace sections in the BBS of super camp.(2)Text preprocessing.Clean text and filter illegal characters according to the contents of crawling data,add military custom dictionary and stop words library to optimize Chinese word segmentation effect,use HANLP named entity recognition tool to identify the name of the organization mentioned in the comment data.(3)Discovery of hot topics.Using the LDA word bag model to clustering the articles according to the topic content and dig out the sensitive topics that peoples care about(4)Emotional analysis.Using the emotional dictionary based approach,rating the comments and statistic peoples public opinion direction(5)Making websites.Using the framework named DJANGO to develop the sword military public opinion network,integrate the above functions into the website.Using “chart.js” and “d3.js” and other data visualization libraries to display various analysis results in the form of charts.