This repository contains a deep learning model developed to classify toxic comments into multiple categories using TensorFlow and Keras. The model employs a Bidirectional LSTM neural network architecture to capture contextual information from comments, resulting in improved classification accuracy.
-
Data Preprocessing: Utilized the TextVectorization layer to preprocess textual data, converting comments into integer sequences for model input.
-
Model Development:
- Implemented a Bidirectional LSTM neural network architecture to capture bidirectional context information from comments.
- Fine-tuned hyperparameters including learning rate, batch size, and dropout rate to optimize model performance.
-
Model Training and Evaluation:
- Trained the model on the training dataset for 2 epochs, leveraging Google Colab for computational resources.
- Evaluated model performance using metrics such as precision, recall, and F1-score on the test dataset.
data/: Contains the dataset used for training and evaluation.notebooks/: Jupyter notebooks detailing the data preprocessing, model development, and evaluation.models/: Saved model files.README.md: Overview of the project.
- TensorFlow 2.x
- Keras
- NumPy
- Pandas
- Matplotlib
- Seaborn