Introduction
- The experiments are seperated into two parts, one is based on the
English
corpus and another is based onChinese
one. - Five different algorithms are deployed to make comparison, one is convolution neural network, and other four are traditional machine learning approaches. Besides, data visualization approaches are used to illustrate the result of diversal models.
- Codes of cnn model is based on NLPproject_6.ipynb by siddarthhair95
Dataset
Chinese corpus
: Data grabbed from Weibo. Chinese_Rumor_DatasetEnglish corpus
: Data are from William Yang Wang, "Liar, Liar Pants on Fire": A New Benchmark Dataset for Fake News Detection. 12.8K short texts are abstracted from PolitiFact.com, which provides detailed analysis report and links to source documents for each case. Liar, liar, pants fire
Environment
Langueage
: Python 3.7IDE
: jupyter notebookPre-trained Word Embedding
: GloVe.6B.100d.txt, sgns.weibo.bigram-char.bz2Database Tookit
: PickleMachine Learning Toolkit
: Scikit-learnDeep Learning Toolkit
: Keras, Tensorflow