📚 Project Title:

Movie Review Sentiment Classification using TensorFlow Hub and IMDB Dataset

---

📝 Project Description:

In this project, we build a simple yet powerful text classification model to predict the sentiment (positive or negative) of movie reviews.
We use the popular IMDB movie review dataset from TensorFlow Datasets and pre-trained word embeddings (GNews Swivel) from TensorFlow Hub to speed up learning and improve performance.

The project involves:

• Loading and preparing the IMDB review data.
• Using GNews Swivel embeddings (20-dimensional vectors) to represent the input text.
• Building a Sequential model with:
  • Pre-trained embedding layer (hub.KerasLayer)
  • A hidden Dense layer with ReLU activation
  • A final Dense layer with Sigmoid activation for binary classification.
• Training the model for 25 epochs to achieve good accuracy.
• Evaluating the model on unseen test data.
• Predicting the sentiment of new movie reviews.

This project demonstrates how transfer learning with pre-trained embeddings can quickly and effectively solve a basic Natural Language Processing (NLP) task without needing a huge custom model.

---

🎯 Key Points:

Component	Details
Dataset	IMDB Reviews (from TensorFlow Datasets)
Embedding	GNews Swivel 20-dimension embeddings (from TensorFlow Hub)
Model Type	Sequential Neural Network
Loss Function	Binary Crossentropy
Optimizer	Adam
Epochs	25
Metrics	Accuracy
Goal	Predict whether a movie review is positive or negative

---

📈 Results:

• After 25 epochs, the model typically achieves ~85–87% validation accuracy.
• Predictions on new examples show good performance in classifying positive vs negative reviews.

---

🚀 Tools Used:

• Python 3
• TensorFlow 2.x
• TensorFlow Datasets (tfds)
• TensorFlow Hub (hub)
• Numpy

---

🧠 Why this Approach?

• Pre-trained embeddings reduce the amount of training data needed and improve generalization.
• Simple model architecture makes it easy to understand and fast to train.
• Transfer Learning is a powerful method for achieving high performance with low effort in NLP tasks.