A Chinese Text Classification Method Based On Lightweight Pre-trained Model and Integrated Data Augmentation

Chinese text classification is an important branch of natural language processing. Currently, the accuracy of Chinese text classification models primarily depends on the effectiveness of large-scale pre-trained models and extensively annotated datasets. However, in practical application scenarios, the scarcity of corresponding data sets and the computational resources required for large-scale pre-trained models pose significant challenges to text classification tasks. To address these challenges, this paper introduces a new Chinese text classification method that utilizes lightweight pre-training and integrated data augmentation techniques to enhance model performance under resource constraints. We evaluated our method on several Chinese text classification tasks, including sentiment analysis and topic classification. The results demonstrate that our approach achieves competitive accuracy compared to state-of-the-art models while significantly reducing the required computational resources. Furthermore, the proposed data augmentation techniques contribute to increased model robustness against overfitting and improved generalization across different domains.