Researches And Implementation Of Technology Event Detection In Social Networks

In contemporary society,social networks have become an important channel for people to pay attention to,obtain and publish information.With the vigorous development of the Internet,the amount of information generated by social networks continues to increase every day.Extracting valuable information from social networks has become an important research direction,and this information will be used in commercial competition,intelligence analysis,public opinion monitoring and other fields.At the same time,more and more scientific institutions and researchers publish information through social networks.This information often contains relevant events in the technology field and is often released earlier than news channels.Therefore,it is also very promising to extract valuable scientific and technological information from social networks by detecting them.In event detection,it is very important to discover events in time,therefore,a good event detection system is very useful.Using technology event detection technology,technology events in social networks can be displayed to users in a concise and intuitive form.This paper studies the event detection algorithm in science and technology field based on thesaurus improvement and designs and implements a hot event detection system in science and technology field by combining data acquisition technology,text representation technology and text classification technology.The main work of this paper is as follows:(1)This paper studies the technology of event detection and based on the characteristics of event detection in the field of science and technology,proposes an improved algorithm of event detection in the field of science and technology,called FSD-improved.The algorithm improves the efficiency of the random hyperplane division of subspaces by combining the thesaurus of science and technology field with local sensitive hash(LSH),thus reducing the number of random hyperplanes and hash tables and improving the efficiency of the algorithm.The experimental results show that when the data set is 2000-150000 documents,the FSDimproved algorithm's processing performance is 10.91%-74.47% higher than Petrovi'c's event detection algorithm based on local sensitive hash,while the accuracy of event detection is almost the same.(2)This paper designs and implements a hot event detection system in the field of science and technology.The system classifies social network data through text classification algorithm,detects hot events using FSD-improved algorithm for classified related data in scientific and technological fields,and displays hot events in a web system based on Django framework.By adopting modular,high availability,and extensible architecture model,the system is divided into data acquisition module,text classification module,storage module,display module,which has better performance and good user experience,and meets the needs of event detection in the field of science and technology.