Multilingual Power and Ideology Identification in the Parliament

本数据集名为‘Multilingual Power and Ideology Identification in the Parliament’，由斯洛文尼亚约瑟夫·斯蒂芬研究所创建，基于ParlaMint语料库，包含来自29个国家和地区的议会演讲转录文本。数据集大小超过770万条演讲，涵盖2015至2022年，用于分析政治倾向和权力地位。创建过程中，通过减少协变量影响和鼓励模型泛化来优化数据集。该数据集适用于多语言和跨文化研究，旨在通过计算方法深入理解议会辩论中的意识形态和权力动态。

This dataset, titled *Multilingual Power and Ideology Identification in the Parliament*, was created by the Jožef Stefan Institute in Slovenia. It is built upon the ParlaMint corpus and includes transcribed parliamentary speech texts from 29 countries and regions. Comprising over 7.7 million speech entries spanning the period from 2015 to 2022, the dataset is designed for the analysis of political orientations and power statuses. During its development, the dataset was optimized by reducing the impact of covariates and promoting model generalization. This resource is applicable to multilingual and cross-cultural research, aiming to gain in-depth insights into the ideological and power dynamics within parliamentary debates through computational methods.