tweets-reliability-prediction

This project is about checking the reliability and unreliability of the Tweets. This project predicts the tweets as reliable and unreliable using 5 deep learning methods i.e. CNN, LSTM, GRU, Transformers (BERT & AlBERT).

Requirments

1. Python 3.7 or greater
2. Tensorflow 2.0 or newer versions
3. scikit-learn 1.1.0 or newer versions
4. pandas
5. matplotlib 3.5 or newer versions
6. numpy 1.17 or newer versions

Installing

To begin with you may need a better machine that can easily can perform neural networks calculation.To start, you can create a new environment. Download the require packages. You can download the require packages using the command below

• Packages

    pip install tensorflow pandas tensorflow-text tensorflow-hub scikit-learn matplotlib numpy

Dataset

The project needs data to generate predictions. To download the dataset visit the following link. The projecct analyze, clean and pre-process the data itself. Also you can preprocess the data as you want.

• Download dataset - Google Drive

Pre-Trained Models

The project also have the pretrained models trained on the dataset. You can use those pre-trained models to fine tune them and can further train on your own dataset. To download the models click the below link

• Download models - Google Drive