Integrated Text Classification Model for Vertical Fields
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing text classification models exhibit poor performance and lack robustness when applied to vertical fields such as medicine, law, and science, due to their limitations in handling short texts and generalization.
Innovation Solution
A method and apparatus for building a text classification model by integrating a probability-based classification model and a similarity-based classification model using a novel loss function, where the models are trained with expert data to improve classification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single classification model is used for text classification, then the model structure is simple and easy to implement, but the classification precision and robustness deteriorate in vertical fields such as medicine, law, and science
Solution Approach 1:
The patent combines multiple classification models (including deep learning models and traditional machine learning models) into an integrated text classification system. The models are trained together with a unified loss function that aggregates individual model loss functions, enabling them to work collaboratively to improve classification precision in vertical fields while maintaining manageable structural complexity through modular architecture design.
2Loss of time
If a single classification model is used for text classification, then the training process is simple and fast, but the generalization performance and robustness worsen
Solution Approach 1:
The patent segments the training process into independent model training units, where each classification model can be trained separately using its own loss function. The unified loss function aggregates these individual loss functions, allowing parallel or sequential training of multiple models without requiring complex joint optimization, thus reducing overall training time while improving generalization performance through diverse model perspectives.
3Ease of manufacture
If existing classification models are applied to short text classification, then the implementation is straightforward, but the classification effect deteriorates
Solution Approach 1:
The patent designs a universal text classification system that can handle both short texts and long texts effectively. The integrated model architecture incorporates components specifically suited for short text classification (such as TF-IDF features and word n-grams) alongside deep learning models, making the system multi-functional and adaptable to different text lengths while maintaining ease of implementation through a unified training framework.
Data Source
AI summary
The present disclosure provides a method and apparatus for building a text classification model, and a text classification method and apparatus. The method of building a text classification model comprises: obtaining a training sample; obtaining a vector matrix corresponding to the text, after performing word segmentation for the text based on an entity dictionary; using the vector matrix corresponding to the text and a class of the text to train a first classification model and a second classification model respectively; during the training process, using a loss function of the first classification model and a loss function of the second classification model to obtain a loss function of the text classification model, and using the loss function of the text classification model to adjust parameters for the first classification model and the second classification model, to obtain the text classification model formed by the first classification model and the second classification model. The text classification method comprises: obtaining a to-be-classified text; obtaining a vector matrix corresponding to the text, after performing word segmentation for the text based on an entity dictionary; inputting the vector matrix into a text classification model, and obtaining a classification result of the text according to output of the text classification model. The text classification effect can be improved through the technical solutions of the present disclosure.