NLP Projects

This repository contains a collection of Natural Language Processing (NLP) projects demonstrating various NLP techniques and models.

Project 1: Sentiment Analysis using Bag of Words

Description: Implemented a sentiment analysis model to classify text reviews as positive or negative using the Bag of Words technique.

Technologies Used:

• Python
• NLTK
• scikit-learn

Key Outcomes/Achievements:

• Achieved F1 score of 0.86 on Amazon reviews dataset
• Demonstrated understanding of text preprocessing techniques (e.g., tokenization, stemming/lemmatization) and feature extraction.

Project 2: Word Embeddings with CBOW and Word2Vec

Description: Implemented and compared CBOW (Continuous Bag of Words) and Word2Vec models to generate word embeddings and capture semantic relationships.

Technologies Used:

• Python
• Gensim
• PyTorch

Key Outcomes/Achievements:

• Gained practical experience with word embedding techniques.
• Visualized semantic relationships using dimensionality reduction techniques (t-SNE).

Project 3: Machine Translation using Transformers

Description: Developed a sequence-to-sequence machine translation model using the Transformer architecture.

Technologies Used:

• Python
• PyTorch
• Transformers library (Hugging Face Transformers)

Key Outcomes/Achievements:

• Demonstrated understanding of attention mechanisms and sequence-to-sequence modeling.

Project 4: Human vs. Machine Text Classification

Description: Developed and compared multiple models (Bag of Words, Multinomial Naive Bayes, BERT, and DistilBERT) to classify text generated by humans versus machines.

Technologies Used:

• Python
• scikit-learn
• Transformers library (Hugging Face Transformers)
• NLTK

Key Outcomes/Achievements:

• Compared the performance of traditional machine learning models with pre-trained language models, demonstrating the performance gains of transformers.
• Achieved 97% accuracy with BERT/DistilBERT.
• Analyzed the strengths and weaknesses of each model.