Analysis of Hot Topics on Microblog Based on Findings and Social Governance
Abstract: In the finding of
hot topics on microblog, the short text, less word, non-standard word use and other
features of microblog have made the traditional identification method of hot
topic powerless. To solve this problem, discovery method of microblog hot topic
based on speed increase has been put forward. First, divide pretreated
microblog according to windows with equal quality, count word frequency of eachword
in each window and express as two-tuple sequence of time; then calculate
increase slope of each word in every two adjacent windows to find words with
faster increasing speed; later calculate increasing speed of users and article
number of microblogs related to the word to make sure whether the word is hot subject
term; finally produce hot topic from the cluster of hot subject term. The
feasibility of this method has been verified by experiment. Experimental
results show that this method has improved identification efficiency and
lowered omission ratio and fall-out ratio, which can effectively and promptly
discover the hot topic of microblog.
Keywords: Topic
identification; Two-tuple sequence of time; Hot topic of microblog; Analysis on
public sentiment
Author: Chan Xu, Rodrigo
Moreno
Journal Code: jptkomputergg160108
