Anlysing performance of Bag-of-Word related features over Hindi Language for Fake News detections. Sequential Models like LSTM and BERT is developed for fake news detection for Hindi Language.
Requirements: Install all the required dependencies mentioned in requirement.txt file.
Dataset: Save all the data set(csv files) in a folder named "final-datasets" in the same folder.
To run BERT Model:
1. Install the dependencies provided in the requirement file.
2. Go to current directory and excute "python BERT.py" command.
After execution:
1. Enter the name of the dataset for which you want to run the model for. first for train then for test. eg, "bbc_ner_train.csv".
2. After feature extraction, model training and prediction, accuracy, class wise F-score and macro-average will be printed on the terminal.
To run LSTM Model:
Following are the instruction needed to be followed on all the datasets
1. Install the dependencies provided in the requirement file.
2. Go to current directory and excute "python LSTM.py" command.
After Execution:
1. Enter the name of the dataset for which you want to run the model for. first for train then for test. eg, "bbc_ner_train.csv".
2. Trained model will be generated.