Classification model training method, classification method, device and equipment

Abstract

The embodiment of the invention provides a classification model training method, a classification method, a device and equipment. The method comprises the following steps: continuously training an original sub-classification model by adopting a sub-classification sample until the difference between a predicted probability value of a sub-class and a sub-classification result in the sub-classification sample meets a first training cut-off condition to obtain a sub-classification model; continuously training the original total classification model by adopting the total classification sample untilthe difference between the predicted probability value of the total category and the total classification result in the total classification sample meets a second training cut-off condition to obtaina total classification model; and continuously training the original multi-classification model by adopting the multi-classification sample until the difference between the predicted probability value of the sub-class and the sub-classification result in the multi-classification sample and the difference between the predicted probability value of the total class and the total classification result in the multi-classification sample meet a third training cut-off condition to obtain the multi-classification model. Through the embodiment of the invention, the classification accuracy of the to-be-classified information can be improved.

Description

technical field [0001] The invention relates to classification technology in the field of artificial intelligence, in particular to a classification model training method, classification method, device and equipment. Background technique [0002] Information classification includes single-label classification and multi-label classification, wherein single-label classification refers to the classification in which the information to be classified corresponds to only one category label, and multi-label classification refers to the classification in which the information to be classified corresponds to multiple different category labels at the same time. Since information to be classified generally has multiple semantics, multi-label classification of the information to be classified can dig out more semantic information corresponding to the information to be classified, and the classification granularity is fine, and the classification effect is good. [0003] When multi-label...

AI Technical Summary

Problems solved by technology

[0003] When multi-label classification is performed on the information to be classified, feature extraction is usually performed on the information to be classified first, and then based on the extracted features and the trained multi-classification model, the probability value of the information to be classified belongs to each category is determined, and finally the probability values ​​are combined with The corresponding preset probability values ​​are compared to determine the category of the information to be classified; however, in the above-mentioned multi-label classification process, if the trained multi-classification model has incomplete multi-label categories or wrong sample labels during the training process, it will As a result, the classification of the trained multi-classification model is inaccurate, so the accuracy of the classification of the information to be classified is low

Images