main data

OverviewThis folder contains datasets and experimental results used in a research project on rumor generation, detection, and debunking. The core data was generated by two large language models—DeepSeek-R1 and qwq-32b—with additional detection results from DeepSeek-V3. The folder includes both direct model outputs and results derived from further analyses based on these outputs. The data is organized into several subfolders, each focusing on specific aspects of the research. Details of the analysis procedures are described in the accompanying manuscript.Folder Structure1. <b>deepseek-r1-debunking</b>This folder contains the results generated by the <b>DeepSeek-R1</b> model for debunking rumors. The files include:<b>R_readability_results.json</b>: Contains readability analysis results for the generated debunking texts.<b>sentiment_analysis_R.json</b>: Contains sentiment analysis results for the generated debunking texts.<b>R_debunking_texts.json</b>: Contains the debunking texts generated by the model.<b>R_debunking_texts_with_similarity.json</b>: Contains the debunking texts along with their similarity scores to the offical debunking texts.2. <b>deepseek-r1-detection</b>This folder contains the results of <b>DeepSeek-R1</b>'s detection of rumors in the <b>FakeNewsNet</b> and <b>Twitter1516</b> datasets. The files include:<b>DR1_detection_twitter1516.json</b>: Detection results for the <b>Twitter1516</b> dataset.<b>DR1_detection_fakenews.json</b>: Detection results for the <b>FakeNewsNet</b> dataset.3. <b>deepseek-r1-generation</b>This folder includes the generated rumors based on specific themes using the <b>DeepSeek-R1</b> model. The themes and corresponding files include:<b>entertainment.json</b>: Rumors generated on entertainment-related topics.<b>financial.json</b>: Rumors generated on financial-related topics.<b>health.json</b>: Rumors generated on health-related topics.<b>disaster-related.json</b>: Rumors generated on disaster-related topics.4. <b>deepseek-v3-detection</b>This folder contains the rumor detection results for the <b>FakeNewsNet</b> and <b>Twitter1516</b> datasets, generated by the updated <b>DeepSeek-V3</b> model. The files include:<b>v3_results_fakenews.json</b>: Detection results for the <b>FakeNewsNet</b> dataset.<b>v3_results_twitter1516.json</b>: Detection results for the <b>Twitter1516</b> dataset.5. <b>qwq-32b-debunking</b>This folder contains the results of the <b>qwq-32b</b> model for debunking rumors. The files include:<b>Q_debunking_texts_with_similarity.json</b>: Contains the debunking texts with similarity scores to the original content.<b>Q_sentiment_analysis.json</b>: Contains sentiment analysis results for the generated debunking texts.<b>Q_debunking_readability_results.json</b>: Contains readability analysis results for the generated debunking texts.<b>Q_debunking_texts.json</b>: Contains the debunking texts generated by the model.6. <b>qwq-32b-detection</b>This folder includes the detection results for <b>FakeNewsNet</b> and <b>Twitter1516</b> datasets, generated by the <b>qwq-32b</b> model. The files include:<b>Q_rumor_detection_results_fakenews.json</b>: Detection results for the <b>FakeNewsNet</b> dataset.<b>Q_rumor_detection_results_twitter1516.json</b>: Detection results for the <b>Twitter1516</b> dataset.7. <b>qwq-32b-generation</b>This folder contains the generated rumors based on specific themes using the <b>qwq-32b</b> model. The themes and corresponding files include:<b>entertainment.json</b>: Rumors generated on entertainment-related topics.<b>financial.json</b>: Rumors generated on financial-related topics.<b>health.json</b>: Rumors generated on health-related topics.<b>disaster.json</b>: Rumors generated on disaster-related topics.Data DescriptionThe following datasets were used in this research:<b>FakeNewsNet</b>: A widely used dataset consisting of fake news stories, which is employed for training and evaluating rumor detection models. This dataset includes news articles labeled as "fake" or "real," and is used in the detection phase of this study.<b>Twitter1516</b>: A dataset containing rumors and non-rumors from Twitter. It is used to evaluate both rumor detection and generation models. The dataset contains tweets labeled as either rumors or non-rumors, providing a benchmark for evaluating the performance of detection models.Both datasets are publicly available and were used to train, test, and evaluate the models in this study. Please refer to the original dataset publications for detailed information on their structure and labeling.<br><br>### 概述 本文件夹包含一项针对谣言生成、检测与辟谣的研究项目所用的数据集与实验结果。核心数据由两款大语言模型（Large Language Model）——DeepSeek-R1与qwq-32b生成，额外包含DeepSeek-V3生成的检测结果。本文件夹同时收录模型的直接输出结果，以及基于这些输出开展进一步分析得到的衍生结果。数据按多个子文件夹进行组织，每个子文件夹聚焦研究的特定方向。分析流程的详细说明见随附的研究手稿。 ### 文件夹结构 1. **deepseek-r1-debunking** 本文件夹包含DeepSeek-R1模型生成的谣言辟谣结果，包含以下文件： - **R_readability_results.json**：收录生成的辟谣文本的可读性分析结果 - **sentiment_analysis_R.json**：收录生成的辟谣文本的情感分析结果 - **R_debunking_texts.json**：收录模型生成的辟谣文本 - **R_debunking_texts_with_similarity.json**：收录模型生成的辟谣文本，以及其与官方辟谣文本的相似度评分 2. **deepseek-r1-detection** 本文件夹包含DeepSeek-R1模型在FakeNewsNet与Twitter1516数据集上的谣言检测结果，包含以下文件： - **DR1_detection_twitter1516.json**：Twitter1516数据集上的检测结果 - **DR1_detection_fakenews.json**：FakeNewsNet数据集上的检测结果 3. **deepseek-r1-generation** 本文件夹包含使用DeepSeek-R1模型基于特定主题生成的谣言数据，主题与对应文件如下： - **entertainment.json**：娱乐主题生成的谣言数据 - **financial.json**：金融主题生成的谣言数据 - **health.json**：健康主题生成的谣言数据 - **disaster-related.json**：灾害相关主题生成的谣言数据 4. **deepseek-v3-detection** 本文件夹包含由升级后的DeepSeek-V3模型在FakeNewsNet与Twitter1516数据集上生成的谣言检测结果，包含以下文件： - **v3_results_fakenews.json**：FakeNewsNet数据集上的检测结果 - **v3_results_twitter1516.json**：Twitter1516数据集上的检测结果 5. **qwq-32b-debunking** 本文件夹包含qwq-32b模型生成的谣言辟谣结果，包含以下文件： - **Q_debunking_texts_with_similarity.json**：收录生成的辟谣文本，以及其与原始内容的相似度评分 - **Q_sentiment_analysis.json**：收录生成的辟谣文本的情感分析结果 - **Q_debunking_readability_results.json**：收录生成的辟谣文本的可读性分析结果 - **Q_debunking_texts.json**：收录模型生成的辟谣文本 6. **qwq-32b-detection** 本文件夹包含qwq-32b模型在FakeNewsNet与Twitter1516数据集上生成的谣言检测结果，包含以下文件： - **Q_rumor_detection_results_fakenews.json**：FakeNewsNet数据集上的检测结果 - **Q_rumor_detection_results_twitter1516.json**：Twitter1516数据集上的检测结果 7. **qwq-32b-generation** 本文件夹包含使用qwq-32b模型基于特定主题生成的谣言数据，主题与对应文件如下： - **entertainment.json**：娱乐主题生成的谣言数据 - **financial.json**：金融主题生成的谣言数据 - **health.json**：健康主题生成的谣言数据 - **disaster.json**：灾害主题生成的谣言数据 ### 数据说明 本研究使用了以下数据集： - **FakeNewsNet**：一款广泛应用的假新闻数据集，包含虚构新闻内容，用于训练与评估谣言检测模型。该数据集包含标注为"虚假"或"真实"的新闻稿件，为本研究的检测阶段所用。 - **Twitter1516**：一款包含Twitter平台谣言与非谣言内容的数据集，用于评估谣言检测与生成模型。该数据集包含标注为"谣言"或"非谣言"的推文，为检测模型的性能评估提供了基准测试集。 两款数据集均为公开可用，本研究使用它们完成模型的训练、测试与评估。如需了解数据集的结构与标注规则的详细信息，请参考数据集的原始发表文献。