Intent Classification with Neural Networks is an NLP project that uses Long Short-Term Memory (LSTM) networks to classify user queries into predefined categories.
- Utilizes GloVe embeddings for high-quality word representations.
- Employs LSTM networks to capture long-term dependencies in text data.
- Offers a detailed pipeline from text preprocessing to model evaluation.
- Includes multiple model configurations to explore the impact of hyperparameters.
- Clone the repository.
- Install the required Python libraries.
- Download and set up the GloVe embeddings.
- Prepare the dataset by running the preprocessing scripts.
The project is tested on a publicly available intent classification dataset, structured with text inputs and intent labels.
- The model training process involves multiple steps including data preprocessing, feature extraction, and training LSTM models.
- Various configurations with different hyperparameters (like hidden dimensions) are tested to find the best performing model.
- Evaluation metrics such as accuracy, precision, recall, and F1-score are calculated to assess the model performance.
- The project includes detailed analysis of the model performance, showcasing the effectiveness of LSTM models in handling text classification tasks.
- Visualizations like confusion matrices are provided to give insights into model predictions.
The project is open-sourced under the MIT License.
- Thanks to the Stanford NLP Group for providing the GloVe embeddings.
- The intent classification dataset contributors for providing a rich dataset for analysis.
For more details, visit the GitHub repository.