Intention analysis method and device, display terminal and computer readable storage medium

Abstract

The invention discloses an intention analysis method and device based on deep learning, a display terminal and a computer readable storage medium. The method comprises the steps that statement sampleswith preset first numerical values are distributed to preset intentions in a classified mode; Selecting statement pairs from the statement samples with the same intention classification to generate positive samples; Selecting statement pairs from the statement samples with different intention classifications to generate negative samples; Training a preset intention analysis model on the basis ofa training set to generate a deep network model, and inputting the positive sample and the negative sample into the deep network model to obtain similarities respectively tending to 1 and 0; When a to-be-identified text is received, inputting the to-be-identified text and the application statement samples in the intention classifications into an intention analysis model as comparison texts; And ifthe target application statement sample exists, determining that the intention of the to-be-identified text belongs to the target intention classification. According to the invention, the generalization ability of the intention analysis model and the transportability of the intention analysis model for different task requirements are enhanced.

Description

technical field [0001] The present invention relates to the technical field of deep learning algorithms, in particular to a deep learning-based intent analysis method, device, display terminal and computer-readable storage medium. Background technique [0002] With the development of science and technology, smart TVs have been widely popularized. Users input voice to smart TVs to search for movies, music and chat. Smart TVs can identify users' intentions through user dialogues. If the user's intentions match the business of smart TVs , then smart TVs can carry out accurate product recommendations. For example, in online advertising scenarios, smart TVs can select advertising pages based on user dialogue to make accurate advertising recommendations. [0003] Intent recognition generally applies short text matching, which plays a vital role in natural language processing tasks such as information retrieval, question answering, and dialogue systems. Conventional text matching ...

AI Technical Summary

Problems solved by technology

[0005] The main purpose of the present invention is to provide a method, device, display terminal, and computer-readable storage medium for intent recognition and analysis based on deep learning, aiming to solve the technical problems of lack of generalization ability and portability in existing intent recognition modules

Images