文件列表:

LICENSE (1070, 2022-06-08)
data/ (0, 2022-06-08)
data/news_cleaned.csv (12978117, 2022-06-08)
results/ (0, 2022-06-08)
results/5gram_200size.txt (1098, 2022-06-08)
results/Batyr_report.pdf (63791, 2022-06-08)
results/generated_text_rnn_chars.txt (9218, 2022-06-08)
src/ (0, 2022-06-08)
src/data_processing.py (1783, 2022-06-08)
src/model_lstm_chars.py (2840, 2022-06-08)
src/model_ngrams.py (4496, 2022-06-08)
src/requirements.txt (180, 2022-06-08)

# News generation Fake news generation using Markov chains (n-gram language modeling) and long-short term memory network (LSTM). ### Usage Install dependencies (keras, tensorflow, numpy, pandas): ```bash # Upgrade your pip3 python3 -m pip install --upgrade pip # Install deps pip3 install -U -r src/requirements.txt ``` Then, clean up your CSV containing tweets (e.g., remove stopwords): ```bash python3 src/data_processing ``` If you want to run n-gram model, run the following command with the output csv file from the previous one: ```bash python3 src/model_ngrams.py ``` If you would like to run LSTM model based on characters, run the following the same way: ```bash python3 src/model_lstm_chars.py ``` These will either output results to command line, or generate a file called `generated_text_rnn_chars.txt`. Feel free to modify, play with the code, and contribute if you find a bug. ### Project report To see the report, click [here](https://github.com/oneturkmen/news-generation/blob/master/./results/Batyr_report.pdf). ### Data clean-up Code [here](https://github.com/oneturkmen/news-generation/blob/master/./src/data_processing.py) ### Examples of generated text Using [n-grams (markov chain)](https://github.com/oneturkmen/news-generation/blob/master/./results/5gram_200size.txt) Using [LSTM](https://github.com/oneturkmen/news-generation/blob/master/./results/generated_text_rnn_chars.txt)

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：