Tag hierarchies generated from social tagging systems can be used in various fields.Current research works mainly focus on the detection of tag relations

but pay much less attention on how to generate high quality tag hierarchy with these relations.This paper studies relation recognition and hierarchy composition methods for Web 2.0 tag hierarchy construction.By analyzing and recognizing the types of the tag relation detected by available methods

quality of tag relation is improved.Moreover

a semantic flow analysis based hierarchy composition method is proposed to acquire tag hierarchy with higher quality.Experimental results measured by various evaluation metrics have shown that by applying relation recognition and semantic flow analysis based hierarchy composition

hierarchies with higher quality with respect to other evaluated methods can be yielded.