artificial intelligence; machine learning; natural language processing; text classification;

In this paper, two datasets for text classification were primarily used in the experiments: AG News and IMDB. The AG News dataset is a widely used four-class news dataset, including four categories: World News, Sports News, Business News, and Technology News. The dataset contains a total of 120,000 samples, with 114,000 samples in the training set and the remaining 6,000 samples in the test set. The IMDB dataset is a movie review dataset used for sentiment analysis, primarily for binary classification tasks, i.e., positive and negative reviews.