MultiBanFakeDetect: An Extensive Benchmark Dataset for Multimodal Under-Resource Bangla Fake News Detection
收藏Mendeley Data2026-04-18 收录
下载链接：
https://data.mendeley.com/datasets/k5pbz9795f
下载链接
链接失效反馈官方服务：
资源简介：
The MultiBanFakeDetect dataset consists of a total of 9,600 text-image instances from online forums, news websites, and social media. Covering political, social, Technology, and entertainment themes, the dataset offers balanced real and fake instances. The data distribution within the labels includes 7,680 instances for training, 960 instances for testing, and 960 instances for validation.
Statistical Overview of Text-Image Pair Data Across Different Types of Fake News
=========================================
Types Training Testing Validation
=========================================
Misinformation 1288 161 162
Rumor 1215 152 151
Clickbait 1337 167 167
Non-fake 3840 480 480
=========================================
Total 7680 960 960
Distribution of Text-Image Pair Data within Labels
=====================================
Label Training Testing Validation
=====================================
1 (Fake) 3840 480 480
0 (Non-Fake) 3840 480 480
=====================================
Total 7680 960 960
Statistical Overview of Text-Image Pair Data Across Different Categories of Fake News
=========================================
Category Training Testing Validation
=========================================
Entertainment 640 80 80
Sports 640 80 80
Technology 640 80 80
National 640 80 80
Lifestyle 640 80 80
Politics 640 80 80
Education 640 80 80
International 640 80 80
Crime 640 80 80
Finance 640 80 80
Business 640 80 80
Miscellaneous 640 80 80
=========================================
Total 7680 960 960
创建时间：
2024-08-08
