fake-news-detection
所属分类：特征抽取
开发工具：Jupyter Notebook
文件大小：5377KB
下载次数：0
上传日期：2022-12-31 12:52:37
上 传 者：
sh-1993
说明： 印尼假新闻检测使用word2vec-lsm和flask。
(Indonesian fake news detection using word2vec-lstm and flask.)
文件列表:
TA_hoax_detection_id.ipynb (653046, 2022-12-31)
app.py (2336, 2022-12-31)
combined_slang_words.txt (21192, 2022-12-31)
dataset.csv (14268781, 2022-12-31)
scraping (0, 2022-12-31)
scraping\cnn crawl viral news.ipynb (14187, 2022-12-31)
scraping\cnn news covid crawl.ipynb (18414, 2022-12-31)
scraping\covid turnbackhoax crawler.ipynb (11793, 2022-12-31)
scraping\data hoax covid crawler.ipynb (11479, 2022-12-31)
scraping\detik crawl viral news.ipynb (12982, 2022-12-31)
scraping\detik news covid crawl.ipynb (27697, 2022-12-31)
scraping\viral turnbackhoaxid crawl.ipynb (8545, 2022-12-31)
templates (0, 2022-12-31)
templates\about.html (224, 2022-12-31)
templates\index.html (677, 2022-12-31)
templates\layout.html (2719, 2022-12-31)
templates\results.html (652, 2022-12-31)
train_w2v.ipynb (11886, 2022-12-31)
# Fake News Detection using Deep Learning
This project proposed Word2Vec-LSTM methods for detecting fake news in Indonesian text language. As many as 4,800 news data from three Indonesian news sites with high credibility (cnnindonesia, detik.co.id, and turnbackhoax.id) are obtained through the crawl/scrap data process and used for this learning project. The average of the 10-fold cross-validation score results shows that:
* Accuracy: 89.42%
* Precision: 88.76%
* Recall: 92.53%
* F1 Score: 89.97%
## Model
link : https://drive.google.com/file/d/1D22KQp4lMTtF4EamzuDkR86lGUbg5seX/
## How to use
1. Clone this repository.
2. Download the model from gdrive link.
3. Run `app.py`, go to the `http://127.0.0.1:5000/`
4. Paste article text in text area input and submit.
- The UI design template reference is by Pooja Bhagat: [https://github.com/581-pooja ](https://github.com/581-pooja/Fake-News-Classification-App/tree/master/templates)
## Note
- This project is only for learning purposes and doesn't work well in real time because was trained on historical and limited data.
## Paper of Project
[1]
R. Yusuf and S. Suyanto, "Hoax Detection on Indonesian Text using Long Short-Term Memory," 2022 5th International Conference on Information and Communications Technology (ICOIACT), 2022, pp. 268-271, doi: 10.1109/ICOIACT55506.2022.9972086.
近期下载者：
相关文件：
收藏者：