Statement classification model training method, statement processing method and equipment

Abstract

The invention provides a statement classification model training method, a statement processing method and equipment, and relates to the technical field of computers. According to the method and device disclosed by the embodiment of the invention, on one hand, the word vectors of partial words in the word vector information corresponding to the sample statements are changed, and the changed word vector information is used as a basis for classifying the sample statements, so that the statement classification model can still accurately classify the changed sample statements, the adaptability and the resistance of the statement classification model to the uncertain change of the statement are enhanced, and the robustness of the statement classification model is improved. On the other hand, the focuses of the statement classification model can be changed by changing the word vectors of part of words in the sample statements, and the statement classification model can accurately classify the sample statements in the training process, so that the statement classification model pays attention to global features and does not excessively pay attention to local features, the situation of over-fitting is avoided, and the prediction accuracy of the statement classification model is improved.

Description

technical field [0001] The disclosure relates to the field of computer technology, in particular to a sentence classification model training method, a sentence processing method and equipment. Background technique [0002] With the development of computer technology, sentences are processed through classification models, and using models instead of manual operations can greatly reduce manual operations and improve data processing efficiency. The classification model usually needs to be trained using sample data, and the classification model is a sentence classification model. Wherein, in a natural language processing scenario, the sentence classification model may be trained so that the sentence classification model can process the sentence. [0003] In related technologies, the sentence classification model training method is usually: obtain sample data, for example, the sample data is a news headline, directly process the sample data through the sentence classification mo...

AI Technical Summary

Problems solved by technology

[0004] However, in real business scenarios, there is no absolutely clean sample data. Among them, clean data refers to correct data with no repeated information and no errors in the data, and unclean data (that is, dirty data) may contain There is duplicate information, or there is an error

The data generated in real business scenarios may include clean data and unclean data. Using sample data that is not absolutely clean to train the model is prone to over-fitting the training data, making the trained model more suitable for processing these sample data. The effect is very good, but it cannot be accurately predicted for data different from these sample data, so there is a problem of poor online performance. Therefore, the prediction accuracy and robustness of the sentence classification model obtained by the above training method are poor

Images