Relation-Extraction-in-NLP-using-Deep-Learning-Models

This project provides two approaches for relation extraction classification using deep learning and machine learning techniques.

Project Overview

• Approach 1: Deep Learning using Neural Networks
• Approach 2: Machine Learning using Support Vector Machines (SVM) with Stacking Classifier

Both approaches are implemented in Python and can be executed on Google Colab.

Setup and Usage in Google Colab

1. Open the Python Notebooks

You can open the provided Python notebooks in Google Colab for training and testing the models:

• Final_Neural_Networks_Model.pynb
• Final_StackingClassifier_Model.ipynb

2. Required Steps

Once you open the Colab notebooks, follow these steps:

2.1 Install Dependencies

Execute the cells one by one to install the required libraries and load the necessary functions.

2.2 Upload Necessary Files

• glove.6B.100d.txt for the neural-based approach:

  • Download the GloVe embeddings file from the link below: GloVe Embeddings
  • Upload it in the Colab runtime under the "Files" section.

• classifier.py for both approaches:

  • Upload the classifier.py file to the Colab environment. This file is already included in the project folder. You can either upload it directly or use a cloud storage link.

Model Training and Testing

1. Deep Learning Approach (Neural Networks)

• Enable GPU Acceleration: Ensure you have T4 GPU enabled in Colab for faster training.
• Execute the cells in the Final_Neural_Networks_Model.pynb to train the deep learning model.
• Use the same notebook for testing and user-testing.

2. Machine Learning Approach (SVM with Stacking Classifier)

• Execute the cells in the Final_StackingClassifier_Model.ipynb to train and test the SVM-based model.
• For faster user-testing compilation, you can also use the Final_StackingClassifier_UserTesting.ipynb notebook.

User Testing

After Training:

For classifying new sentences after model training, use the classifier.py module. Follow these import statements for each approach:

• For Deep Learning Approach:

  from classifier import classify_relation

• For SVM Approach:

  from classifier import classify_relation_svm