Shallow layer model and depth model combination-based question text classification method

Abstract

The invention relates to a shallow layer model and depth model combination-based question text classification method, and belongs to the technical field of computer natural language processing. The method comprises the following steps of: firstly extracting a feature word set of a question text, obtaining corresponding feature word weights by utilizing normalized word vectors after vectorization,and taking the corresponding feature word weights as a part of input of a shallow layer linear model; convoluting the question text by a convolutional network by using multiple convolution cores withdifferent window sizes, rearranging the feature vectors extracted by different convolution cores with same-length convolution windows, respectively inputting the feature vectors into corresponding recurrent neural networks, and finally taking syntax semantic feature vectors, obtained by linking outputs of the plurality of recurrent neural networks together, of the question as another part of inputof the shallow layer linear model; and finally obtaining a classification result of the question by the shallow layer model according to an input spliced by the feature word vectors and an output ofa depth model. The method is capable of overcoming the defects of single depth models and effectively enhancing the question classification correctness.

Description

technical field [0001] The invention relates to a question text classification method based on the combination of a shallow model and a deep model, and belongs to the technical field of computer natural language processing. Background technique [0002] Question text classification belongs to short text classification and plays an important role in automatic question answering systems. Question text classification mainly classifies questions by analyzing their content. In the early days, there were rule-based methods, which used the corresponding relationship between keywords or grammatical patterns of questions and question types to classify questions. This method works well for question classification with obvious interrogative words or question category feature words, but not for more complex questions or question sentence texts without obvious category feature words, and the flexibility of the method is not enough. High, the workload is large, and the subjectivity of q...

AI Technical Summary

Problems solved by technology

[0004] The present invention provides a question text classification method based on the combination of a shallow model and a deep model. Aiming at the problems existing when a single deep model faces unbalanced training data, the traditional shallow model has strong memory characteristics for features, effectively improve the accuracy of question classification

Images