radwa-f/LSTM-IMDB-SentimentAnalysis

LSTM IMDB Sentiment Analysis - Baseline Model

Traditional LSTM model for binary sentiment classification achieving 89% accuracy on IMDB movie reviews.

This model serves as a baseline in a comparative study demonstrating the superiority of transfer learning (DeBERTa: 96%) over traditional approaches.

Quick Start

import pickle
from tensorflow import keras
from huggingface_hub import hf_hub_download

# Download
model_path = hf_hub_download("radwa-f/LSTM-IMDB-SentimentAnalysis", "model.h5")
tok_path = hf_hub_download("radwa-f/LSTM-IMDB-SentimentAnalysis", "tokenizer.pkl")

# Load
model = keras.models.load_model(model_path)
with open(tok_path, 'rb') as f:
    tokenizer = pickle.load(f)

# Predict
text = "This movie was amazing!"
seq = tokenizer.texts_to_sequences([text])
padded = keras.preprocessing.sequence.pad_sequences(seq, maxlen=200)
pred = model.predict(padded)[0][0]

print(f"{'POSITIVE' if pred > 0.5 else 'NEGATIVE'} ({pred:.2%})")

Performance Comparison

Model	Accuracy	Type
LSTM (this)	89%	Traditional RNN
DeBERTa	96%	Transfer Learning
Difference	-7%

Architecture

Embedding(vocab_size, 128)
LSTM(128)
Dense(64, relu)
Dropout(0.5)
Dense(1, sigmoid)

Training

• Dataset: IMDB (50K reviews)
• Framework: Keras/TensorFlow
• Loss: Binary Crossentropy
• Optimizer: Adam

Model Creators

Radwa Fattouhi

• Final-year Software Engineering Student
• École Nationale des Sciences Appliquées (ENSA), El Jadida, Morocco
• Email: [email protected]
• LinkedIn: radwa-fattouhi
• GitHub: radwa-f

Amine Boktaya

• Final-year Software Engineering Student
• École Nationale des Sciences Appliquées (ENSA), El Jadida, Morocco
• Email: [email protected]
• LinkedIn: amine-boktaya
• GitHub: BoktayaAmine

Citation

If you use this model in your research, please cite:

@misc{fattouhi2025imdb,
  author = {Fattouhi, Radwa and Boktaya, Amine},
  title = {LSTM IMDB Sentiment Analysis: Transfer Learning vs Traditional Approaches},
  year = {2025},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/radwa-f/DeBERTA-Imdb-SentimentAnalysis}}
  howpublished = {\url{https://huggingface.co/radwa-f/LSTM-IMDB-SentimentAnalysis}}
}

Related Work

Published Research:

• AgriAlertX: Climate-driven disaster prevention for agriculture
• Journal: SoftwareX (Elsevier)
• DOI: 10.1016/j.softx.2025.102350

Related Projects:

• Riot Detection System (DeBERTa for social media classification)
• Tweet Detoxification (BART for style transfer)

Author

Radwa Fattouhi - ENSA El Jadida, Morocco