STUDY ON UNCONVENTIONAL EMERGENCY EVENT CLASSIFICATION BASED ON SUBJECT OBJECT MODEL

摘要

In order to solve the problems on network information classification of unconventional emergency event, a web text classification model based on SOM (Subject Object Model) was proposed. The process of text block extraction algorithm and the generated procedure of the class feature vector set were given, the text classification strategy of multi-subject support vector machine was proposed, construction and storage of SOM was completed. The experimental results show that this classification method is feasible and has higher performance on precision and recall.