Tagging method and apparatus based on structured data set

摘要

Tagging methods and apparatus, including computer program products, based on a structured data set. Classification models are created for respective nodes in the structured data set of an event. Public opinions on the event are acquired. The opinions are tagged to corresponding nodes of the structured data set using the created classification models. The tagging methods and apparatus provide well-ordered, focused public opinions for each event to users, and exhibit the evolution of the public opinions along with time.

主权项

1. A tagging method based on a structured data set, comprising:
receiving a reported real-life news event from an official source; extracting the structured data set from the received reported news event; creating an event tree representing the extracted structured data set of the reported real-life news event, the event tree including a plurality of nodes, wherein each node represents a different feature of the reported news event; creating a plurality of classification models for each node in the event tree; acquiring public opinions on the reported news event by searching the World Wide Web for user-generated contents including the public opinions; and tagging the acquired public opinions to corresponding nodes of the event tree using the plurality of created classification models, wherein the tagging comprises:
extracting feature information of a particular public opinion;inputting the feature information of the particular public opinion into the plurality of classification models of a current node;outputting a classification result which exhibits similarity level between the particular opinion and the current node in form of numerical value; andrecording the current node, and replacing a predetermined threshold value of the similarity level with an outputted value of the similarity level, if the value of the outputted similarity level is larger than the predetermined threshold value.