Analyzing German Parliamentary Speeches: A Machine Learning Approach for Topic and Sentiment Classification

2025-08-05

AI Summary

This study investigates how political discourse in the German Bundestag dynamically evolves as parties transition between governing and opposition roles. Drawing on approximately 28,000 parliamentary speeches from the past five years, we develop and train a dual-task machine learning model: one for topic classification (mean AUROC = 0.94) and another for fine-grained sentiment classification (AUROC = 0.89), both trained on human-annotated data. Results reveal systematic discursive shifts during role transitions: governing parties prioritize governance-related topics and exhibit more restrained sentiment, whereas opposition parties emphasize ideological themes and express heightened negative sentiment. This work provides the first empirical validation that both “governing responsibility” and “ideological positioning” jointly shape political communication strategies. It contributes a reproducible computational framework for comparative political discourse analysis and advances theoretical understanding of institutional role effects on rhetorical behavior.

Technology Category

• Natural Language Processing: Discourse, Pragmatics & Argument Mining
• Machine Learning: Multimodal Learning
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Web Mining and Content Analysis: Sentiment analysis and opinion mining
• Social Networks and Social Media: Media and governance, opinion dynamics, filter bubbles, polarization
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

This study investigates political discourse in the German parliament, the Bundestag, by analyzing approximately 28,000 parliamentary speeches from the last five years. Two machine learning models for topic and sentiment classification were developed and trained on a manually labeled dataset. The models showed strong classification performance, achieving an area under the receiver operating characteristic curve (AUROC) of 0.94 for topic classification (average across topics) and 0.89 for sentiment classification. Both models were applied to assess topic trends and sentiment distributions across political parties and over time. The analysis reveals remarkable relationships between parties and their role in parliament. In particular, a change in style can be observed for parties moving from government to opposition. While ideological positions matter, governing responsibilities also shape discourse. The analysis directly addresses key questions about the evolution of topics, sentiment dynamics, and party-specific discourse strategies in the Bundestag.

Problem

• Classify topics and sentiments in German parliamentary speeches
• Analyze topic trends and sentiment distributions across parties
• Examine discourse strategies evolution in the Bundestag over time

Innovation

• Machine learning models for topic classification
• Sentiment analysis on parliamentary speeches
• AUROC performance metrics evaluation