Chinese novel dialogue character recognition method

Abstract

The invention discloses a Chinese novel dialogue character recognition method, which comprises the following steps of: 1, extracting a plurality of sentences before and after a dialogue to form an associated text, and extracting all character names as candidates; 2, selecting sentences before and after the conversation, judging whether a person name exists in the subject or not, if so, determiningthat the conversation person is the person name in the subject, and if not, continuing the next step; 3, judging whether someone names pronouns in the subject, if so, continuing the step 4, and if not, continuing the step 5; 4, matching the personal pronouns in the subject with the sexes of the candidates, judging whether the number of the candidates with the consistent sexes is 1 or not, if so,determining that the dialogue character is the candidate with the consistent sexes, and if not, continuing the step 5; and 5, calculating a candidate vector of each candidate, inputting the candidatevectors into the MLP network to obtain scores of the candidates, and determining that the dialogue person is the candidate with the highest score. The invention belongs to the technical field of information, and can realize automatic and accurate identification of dialogue characters in a Chinese novel.

Description

technical field [0001] The invention relates to a method for identifying dialogue characters in Chinese novels, which belongs to the technical field of network communication. Background technique [0002] With the development of the mobile Internet, reading Chinese novels has increasingly become an important form of entertainment in people's daily lives. However, people are not only satisfied with the enjoyment brought by reading Chinese novels, but also hope to bring auditory experience to Chinese audio novels. The three elements of speech are information, timbre, and rhythm. In the process of automatic synthesis of audio novels, the information is the text information of the dialogue in the novel. The timbre refers to the characteristics of the speaker itself, and the rhythm refers to the pitch, speed, etc. of the voice. Using NLP correlation The technology recognizes the speaker of the dialogue in the novel, and then can synthesize the relevant timbre according to the sp...

AI Technical Summary

Problems solved by technology

In this technical solution, the production of feature vectors is manually selected, which cannot fully express the semantic information before and after the position of the candidate. The quality of the result depends heavily on the quality of the manually selected feature vectors. The dilemma of being unable to fully represent the semantic and other feature information of candidates

Images