AraDhati+
收藏资源简介:
AraDhati+数据集是一个用于阿拉伯文本数据主观性评估的全面数据集。该数据集通过利用现有的阿拉伯语数据集和集合(ASTD、LABR、HARD和SANAD)开发而成。该数据集为阿拉伯语言模型(XLM-RoBERTa、AraBERT和ArabianGPT)进行微调,以实现有效的主题性分类。此外,还尝试了集成决策方法来利用各个模型的优点。该方法在阿拉伯语主观性分类中取得了97.79%的准确率。结果表明,该方法在解决阿拉伯语资源有限带来的挑战方面非常有效。
The AraDhati+ dataset is a comprehensive resource for subjective assessment of Arabic textual data. It was constructed by leveraging existing Arabic datasets and collections: ASTD, LABR, HARD, and SANAD. This dataset is utilized to fine-tune prominent Arabic language models including XLM-RoBERTa, AraBERT, and ArabianGPT for effective topic classification. Furthermore, ensemble decision strategies were explored to leverage the complementary strengths of individual models. This ensemble approach achieved an accuracy of 97.79% on Arabic subjective classification tasks. The experimental results demonstrate that this method effectively addresses the challenges associated with limited Arabic language resources.




