Document classification method and device and terminal

Abstract

The invention provides a document classification method and device and a terminal, and belongs to the technical field of deep learning. The method comprises the following steps: determining a to-be-classified first document; determining a plurality of second documents corresponding to first option information in a plurality of pieces of option information and a plurality of third documents corresponding to second option information; determining a first prediction vector used for indicating the similarity relationship between the first option information and the second option information and a second prediction vector used for indicating the reasoning relationship between the description information and the first option information according to the description information of the first document, the plurality of the second documents and the plurality of the third documents; determining a third prediction vector of the first document according to the first prediction vector and the second prediction vector; and determining the category of the first document according to the third prediction vector. According to the document classification method and device, the category of the first document is determined by determining the similarity relation and the reasoning relation of the related documents of the first document, and therefore subjective deviation generated when the documents are classified manually is avoided, manpower consumption is reduced, and cost is reduced.

Description

technical field [0001] The present disclosure relates to the technical field of deep learning, and in particular to a document classification method, device and terminal. Background technique [0002] Students and electronic education manufacturers have increasingly strong demands for formulating personalized learning strategy algorithms. When formulating personalized learning strategies, we often recommend topics of different difficulty categories for different students according to their progress or ability. Therefore, when formulating a personalized learning strategy, it is necessary to classify the documents according to the difficulty of the topic. [0003] In related technologies, when classifying documents for a topic, the difficulty of the first document corresponding to the topic is often determined through the method of expert judgment based on the subjective judgment of the expert; or, a sample survey is conducted among students, and the topic is determined throu...

AI Technical Summary

Problems solved by technology

[0005] Embodiments of the present disclosure provide a document classification method, device, and terminal, which are used to solve the problem that in the current method of classifying the first document corresponding to a topic through expert judgment or student feedback, all judgments are made by humans, resulting in prediction There are subjective deviations in the results, and the prediction process requires a lot of manpower and high manpower costs

Images