Detection of Self-Introductions in Legislative Testimony

2026-08-07

AI Summary

This study addresses the task of automatically detecting self-introductions and extracting speaker names in legislative hearing transcripts. The authors construct a large-scale dataset comprising 1.54 million utterances and generate labels using a name-matching heuristic. They propose a hybrid classification framework that integrates multidimensional features—including bag-of-words representations, positional and structural signals, leading phrases, and discourse context—and employ ensemble learning models such as XGBoost. Notably, they introduce fine-tuned BERT probability outputs as complementary features for the first time. Experimental results demonstrate that the XGBoost model achieves an F1 score of 0.9747, which further improves to 0.9782 when incorporating BERT-derived features, reducing the number of test errors from 241 to 207 and significantly outperforming baseline approaches.

Abstract

Self-introductions are common in legislative committee testimonies. Successfully detecting them and extracting the speaker's name is enormously helpful in the task of speaker identification in the context of government meetings. In this paper, we present a pipeline for detection of self-introductions in legislative committee testimony using machine learning. We construct a training dataset from 1.54 million utterances spanning five state legislative sessions, apply a name-matching heuristic to generate automatic labels, and train three classifiers: a decision tree, random forest, and XGBoost to find self-introductions and extract the speaker's name. We construct a feature set combining bag-of-words, positional context, structural signals, introductory phrase indicators, and discourse context features. Among the three classifiers, XGBoost achieves the best performance with an F1 score of 0.9747 and the fewest total errors; adding fine-tuned BERT probability features improves this further. As an extension, we score the full candidate dataset with a fine-tuned BERT classifier and add BERT probability outputs as features. This BERT-augmented XGBoost model improves F1 from 0.9747 to 0.9782 and reduces total test errors from 241 to 207. The primary gain over the decision tree baseline (F1 0.9323) is driven by discourse context features and the boosting ensemble strategy; BERT provides a modest complementary signal. Analysis of false positives reveals that a minority are genuine self-introductions mislabeled due to name inconsistencies in the source data, indicating that measured metrics modestly understate true performance.

Problem

• self-introduction detection
• speaker identification
• legislative testimony
• name extraction
• government meetings

Innovation

• self-introduction detection
• speaker identification
• XGBoost
• BERT augmentation
• legislative testimony