文件列表:

LICENSE (1063, 2020-06-07)
Procfile (23, 2020-06-07)
__init__.py (0, 2020-06-07)
analysis (0, 2020-06-07)
analysis\analyze.py (1941, 2020-06-07)
analysis\pred_results.csv (1352, 2020-06-07)
analysis\true_results.csv (81204, 2020-06-07)
covid19_articles (0, 2020-06-07)
covid19_articles\test_set (0, 2020-06-07)
covid19_articles\test_set\breitbart.csv (1243226, 2020-06-07)
covid19_articles\test_set\cnn.csv (41270, 2020-06-07)
covid19_articles\test_set\combined.csv (2851381, 2020-06-07)
covid19_articles\test_set\latimes.csv (626604, 2020-06-07)
covid19_articles\test_set\newyorker.csv (344293, 2020-06-07)
covid19_articles\test_set\npr.csv (420176, 2020-06-07)
covid19_articles\test_set\nytimes.csv (70823, 2020-06-07)
covid19_articles\test_set\theguardian.csv (442473, 2020-06-07)
covid19_articles\test_set\vox.csv (3147029, 2020-06-07)
covid19_articles\test_set\wired.csv (83025, 2020-06-07)
covid19_articles\val_set (0, 2020-06-07)
covid19_articles\val_set\covid_19_articles.csv (81204, 2020-06-07)
covid19_articles\webscraper.py (3328, 2020-06-07)
data (0, 2020-06-07)
data\Emotion_Affect.raw (163314, 2020-06-07)
data\__init__.py (0, 2020-06-07)
data\dataset.py (3897, 2020-06-07)
data\emotion_affect.py (2469, 2020-06-07)
data\fake_news.py (2378, 2020-06-07)
data\feature_extraction.py (3899, 2020-06-07)
data\news_category.py (2134, 2020-06-07)
data\preprocessing.py (2623, 2020-06-07)
data\stanford_sentiment.py (2549, 2020-06-07)
evaluate.py (6694, 2020-06-07)
experiments.ipynb (51809, 2020-06-07)
historical.ipynb (41156, 2020-06-07)
images (0, 2020-06-07)
images\COVID19_News_Analyser.pdf (1113073, 2020-06-07)
... ...

# covid-news-analyzer Analyze news articles related to covid-19 based on different metrics: 1. Sentiment Analysis 2. Emotion Detection 3. News Article Categorization 4. Fakeness You can access the tool online at: https://covid-news-analyzer.herokuapp.com/ (this doesn't currently support fakeness detection due to memory constraints on Heroku's free tier) ---- ## How To Run Locally Install requirements from `requirements.txt` using `pip install -r requirements.txt` and run: ### Training ```bash python -u main.py --dataset --models --feats --save_path --save_results ``` - can be one of `emo_aff, stan_sent, news_cate, fake_news` representing the standard datasets used to train the models - is a space separated list of models to train - mnb, svm, xgb, ada, rf, lr - is a space separated list of feature transformations to use - bow, tfidf, ngram This would train the specified list of models on all the specified feature transformations for the given dataset, and save the results in a csv file, as well as the performing model on the test portion of the dataset. ## Evaluation Run `python evaluate.py` to obtain the predictions on the COVID-19 article test set, and use `python analysis/analyze.py` to obtain the evaluation scores. ## Web Portal Run `python wsgi.py` to launch the web portal. This lets you analyze a specific article against trained models placed in `output/model_dump` directory. ## Report You can find more details about the project here.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：