Node Selecting Strategy For Hot Events Detecting Subnet In Sina Weibo Microblogging

Microblogging has become a popular means of communication anddissemination of information, a large number of social events are spread inmicroblogging. Detecting hot events in microblogging are becomingincreasingly important. However, the hot event detection in microbloggingfaces some big challenges. Large numbers of active users can produce a lotof microblogs in a short time, which makes it impossible to process all themicroblogs real-time.The paper achieves a hot events detection system on China’s largestmicroblogging service Sina Weibo platform. As the real-time processingof all Weibo produced over a period of time to detect hot events iseconomically feasible, the article adopted a strategy, which is to detect hotevents in limited resource by monitoring the microblogs of a small portionof microblogging users. The main purpose of the paper is to provide subnetselecting algorithm to hot events detection system which achieves hotevents detection by monitoring the subnet node. The paper first proposes theconcept of degree of cover and an algorithm which aims at covering all thesample hot events. For the shortage of the algorithm, the paper thenproposes the concept of hot events participation probability of nodes and aprobability algorithm based on it. Considering the cost of monitoring nodesare different, the paper proposes a concept of cost of nodes and anoptimization algorithm. About525hot events are collected in total.294of them are sample dataset and the rest are test dataset. The proposed threealgorithm are applied on the dataset respectively. The result shows that theoptimization algorithm can detect more hot events with less cost. The hotevents detection rate is70%.