Semantic recognition method and device based on self-updating NLP and medium

The invention discloses a semantic recognition method and device based on self-updating NLP and a medium, and relates to the technical field of semantic recognition, and the method comprises the steps: building a basic semantic recognition model, calculating the question and answer similarity between the nth question and answer and the previous (n-1) th question and answer, carrying out the semantic analysis of the nth question and answer, and updating a dialogue feature library; collecting the accuracy of each question and answer of the first dialogue of the user, calculating to obtain a question and answer semantic recognition accuracy coefficient, screening all questions and answers in the first dialogue, and constructing a first dialogue feature library; generating a user initial semantic recognition model, analyzing the newly added features of the newly added dialogue, updating and iterating the user feature library, and updating and iterating the user semantic recognition model at the same time; calcula