Semantic model method for intelligent interaction

The invention discloses a semantic model method for intelligent interaction. The method comprises the following steps of A, performing work segmentation on information sent by a user; B, recognizing whether words, phases and word groups after the word segmentation in the step A belong to an entity or not; C, performing semantic annotation analysis on the words, the phases and the word groups subjected to the word segmentation in the step A; D, performing text error correction on the words, the phases and the word groups subjected to the word segmentation in the step A; E, performing syntactic analysis on the information sent by the user; F, performing weight correction on the information sent by the user and the words, the phases and the word groups subjected to the word segmentation in the step A. The semantic model method has the advantage that the recognition accuracy on a user instruction is high.