Research And Application Of The Algorithm For Key Nodes Detection From Social Network On The Basis Of Events

With the development of the Internet, social network has been more and more popular and mature, plus the participation of users has reached an unprecedented high level. As social network tends to become the latest source of most news, the detection of new event has been the focus of the academia and government. The social network events are spread by users’ tweeting and retweeting, some of which play significance on the communication as core users, so the detection of core users has always been the hot spots of academia. However, the present new event detection method on core users existing with large amount is in lack of efficiency. To design a detection method which can filter all key nodes to obtain true key ones, plays a role in decreasing the workload about dealing with large amount of users as to improve the efficiency on detecting new event.On the basis of traditional algorithms of new event detection, this paper is to discuss about the existing methods obtaining key nodes and to make a breakthrough by achieving higher efficiency, furtherly to point out the Event Based Key Node Detection (EBKND) algorithm, which can remove repeated nodes while comparing the spreading functions of different nodes so as to decrease largely the amount of key nodes while detecting new event on-line. Also, EBKND can work off-line in the beginning, which means it can save the required time and hardware cost. In a conclusion, the feasibility and effectiveness of the above detection method can be proved through experiments, which ensures its high efficiency and stability while maintaining the acuity while detecting new event.