INTENT ANALYSIS METHOD AND APPARATUS, DISPLAY TERMINAL, AND COMPUTER READABLE STORAGE MEDIUM

摘要

A deep learning-based intent analysis method and apparatus, a display terminal, and a computer readable storage medium. The method comprises: allocating a preset number of sentence samples for various preset intent classifications; selecting sentence pairs from the sentence samples of same intent classifications, and generating positive samples; selecting sentence pairs from the sentence samples of different intent classifications, and generating negative samples; training a preset intent analysis model on the basis of a training set, and generating a deep network model, wherein similarities obtained when the positive samples and the negative samples are inputted into the deep network model respectively tend towards 1 and 0 (S30); upon receiving a text to be identified, inputting the text to be identified and application sentence samples in various intent classifications into the intent analysis model as comparison texts; if a target application sentence sample is present, determining that the intent of the text to be identified belongs to a target intent classification. The present method enhances the generalization ability of the intent analysis model and its portability for different task requirements.