Micro Auto Blogging System by Using Granular Tree-Based Context Model

This paper suggests an automatic blogging system based on context cognition technology considering the context of a user's location and time. This system is modeled by preprocessing and combining user context and using granular tree. This modeled context infers user's behavior by using Naive Bayes Classification and user's destination by using sequence matching technique. Sentences that fit situations are generated and automatically blogged using 4W structure. The evaluation of bloeging sentences shows 85.7% accuracy on average and verifies that the context modeling technique that suggests automatic blogging is effective.