Sentiment Analysis on IMDB Reviews

This repository contains a complete sentiment analysis pipeline applied to the IMDB movie reviews dataset. The project demonstrates end-to-end NLP and deep learning techniques, from data ingestion and preprocessing to model training, evaluation, and export.

---

📂 Project Structure

• sentiment_analysis_notebook.ipynb — Jupyter notebook with the full workflow
• a1_IMDB_Dataset.csv — Raw dataset of 50,000 IMDB reviews (labeled positive/negative)
• README.md — This document

---

🔎 Dataset

The IMDB dataset consists of 50,000 movie reviews evenly split between positive and negative sentiment labels. We load it with pandas for quick exploration.

---

🧹 Preprocessing Steps

1. HTML Tag Removal — Strips away any HTML markup using regex.
2. Lowercasing — Converts all text to lowercase.
3. Regex Cleaning — Removes digits, punctuation, and stray characters with regular expressions.
4. Stopwords Removal — Eliminates common English stopwords for noise reduction.
5. Tokenization — Uses Keras Tokenizer to convert text into sequences of word indices.
6. Sequence Padding — Pads or truncates sequences to a fixed length for uniform model input.

---

🤖 Model Architecture

A deep learning network built with TensorFlow / Keras:

• Embedding Layer — Learns dense vector representations of words.
• LSTM Layer — Captures sequential dependencies in text data.
• Dense Layers — Fully connected layers to interpret features and output a binary sentiment prediction.

Compilation Settings:

• Loss: Binary crossentropy
• Optimizer: Adam optimizer
• Metrics: Accuracy

---

🚀 Training & Evaluation

• Training — The model is trained using model.fit() on the tokenized and padded review sequences.
• Evaluation — Performance is measured on a held-out test split via model.evaluate().

Typical results include accuracy scores and loss curves, easily visualized within the notebook.

---

⚙️ Requirements

• Python 3.8+
• pandas
• numpy
• nltk
• tensorflow
• scikit-learn
• joblib
• tqdm

Install dependencies with:

pip install -r requirements.txt

---

📈 How to Run

1. Clone the repository:

   git clone https://github.com/hanzlahyasir/nlp_with_nn.git
   cd nlp_with_nn

2. Install requirements:

   pip install -r requirements.txt

3. Launch the notebook

4. Follow the cells to preprocess data, train the model, and evaluate performance.

---

🏷️ Key Techniques & Libraries

• pandas, numpy — Data handling and numerical operations
• nltk — Stopwords and basic text utilities
• re (regex) — Text cleaning
• TensorFlow / Keras — Deep learning framework
• joblib — Parallel preprocessing
• tqdm — Progress bars

---

📂 Next Steps

• Experiment with different RNN variants (e.g., Bidirectional LSTM)
• Tune hyperparameters (embedding size, sequence length, learning rate)
• Add advanced evaluation metrics (ROC AUC, confusion matrix)
• Deploy the model via a simple web API for real-time sentiment analysis

---

Prepared for showcasing NLP and deep learning skills to hiring teams.