The Tree Model of Discourse Information proposed by Professor Du Jinbang can be employed to transform a discourse into a tree of discourse information knots. The discourse tree can give researchers detailed information of a discourse, including the macro-structure and the micro-components. It can be employed for authorship attribution, forensic speaker recognition, forensic psycholinguistic analysis, legal English teaching research and etc. For example, by combining adjacent discourse information knots, Sunbo (2013) constructs discourse information bigrams. He proves that discourse information bigrams can be applied to authorship attribution with a quite good discriminating potential. Through discriminant analysis, discourse information bigrams show both inter-author variation and intra-author constancy (please see the following figure).