From Promises to Accountability: Comparative AI Modeling and Visualization of Electoral Commitments in Nepali Political Texts

📄 arXiv: 2610.32150v1 📥 PDF

作者: Prakash Poudyal, Nischal Karki, Amrit Dahal, Rajani Chulyadyo, Sameer Tamrakar, Bal Krishna Bal

分类: cs.CL

发布日期: 2026-10-23

备注: Under review at Natural Language Processing (Cambridge University Press); revision submitted. 34 pages, 5 figures, 13 tables

💡 一句话要点

构建选举承诺挖掘系统以提升透明度与问责制

🎯 匹配领域: 支柱九：具身大模型 (Embodied Foundation Models)

关键词: 选举承诺挖掘 机器学习 深度学习 数据集构建 政治透明度 公民参与 问责制

📋 核心要点

  1. 现有方法未能系统性地归档选举承诺，导致选民难以追责。
  2. 研究开发了一种新系统，能够结构化候选人在选举期间的承诺，提升透明度。
  3. 实验中，LaBSE-LR和fastText-LSTM模型表现优异，准确率分别达到0.897和0.914。

📝 摘要（中文）

选举承诺挖掘是分析候选人或政党在选举期间所做承诺的过程。然而，许多南亚发展中国家的候选人未能兑现其承诺，导致选民难以追责。本文旨在开发一个系统，结构化候选人在选举期间的承诺。研究创建了首个尼泊尔政治文本的金标准数据集，并评估了多种机器学习和深度学习模型，最终结果集成到移动应用中，以增强透明度和公民参与。

🔬 方法详解

问题定义：本研究解决的是南亚国家选举承诺未兑现的问题，现有方法缺乏系统性和透明度，导致选民无法有效追责。

核心思路：通过构建一个结构化系统，自动提取和分析候选人的选举承诺，提升承诺的可追溯性和透明度。

技术框架：整体架构包括数据收集、数据标注、模型训练和应用开发四个主要模块。首先收集尼泊尔政治文本，然后进行标注，接着训练多种机器学习和深度学习模型，最后将结果集成到移动应用中。

关键创新：本研究创建了首个尼泊尔政治文本的金标准数据集，并在此基础上评估多种模型，显著提升了选举承诺的提取和分析能力。

关键设计：在模型训练中，使用了Cohen's kappa作为评估标准，LaBSE-LR和fastText-LSTM模型的损失函数和网络结构经过优化，确保了高准确率和F1-score。具体参数设置和模型架构在实验中进行了详细调优。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果显示，LaBSE-LR模型在机器学习中取得了0.897的准确率和0.947的AUROC，而fastText-LSTM在深度学习中表现最佳，准确率达到0.914。Qwen3-8B在少样本设置下的宏F1-score为0.848，展示了模型在实际应用中的有效性。

🎯 应用场景

该研究的潜在应用领域包括选举监测、政治透明度提升和公民参与增强。通过移动应用，用户可以实时查看候选人的承诺，促进选民对政治过程的关注和参与，进而推动民主问责制的发展。

📄 摘要（原文）

Election Commitment Mining refers to the analysis of election commitments made by electoral candidates or political parties during election campaigns. However, many elected candidates and political parties, especially in developing South Asian countries, do not fully deliver on the promises they make before elections. Even after the elections, commitments presented through election manifestos, posters, pamphlets, videos, and other campaign materials are often not systematically or transparently archived. As a result, these promises tend to fade away over time, making it difficult to hold the elected candidates accountable for their commitments. To address this persistent issue, this research aims to develop a system capable of structuring the commitments made by candidates during election periods. In this work, we created the first gold-standard dataset of Nepali political texts (Cohen's kappa = 0.777) and evaluated machine learning, deep learning and large language models to automatically extract and analyze election commitments. LaBSE-LR achieved the best results among machine learning models (0.897 accuracy, 0.896 macro F1-score, 0.947 AUROC), while fastText-LSTM achieved the highest performance among deep learning models (0.914 accuracy, 0.913 macro F1-score, 0.961 AUROC). Among the three evaluated large language models, Qwen3-8B performed best in the few-shot setting (0.848 macro F1-score). The outcome of this work is integrated into a mobile application, where commitments can be visualized with the support of GAI, enhancing transparency, civic engagement, and electoral accountability.