Pandiri Lavanya / Suneel Kumar Duvvuri
Librería Samer Atenea
Kálamo Books
Librería Elías (Asturias)
Librería Kolima (Madrid)
Librería Proteo (Málaga)
The rapid growth of digital platforms has led to a huge increase in user-generated text, especially movie reviews. Manually analyzing such data is difficult and time-consuming. Sentiment Analysis, an important area of Natural Language Processing (NLP), helps identify opinions expressed in text automatically. This study develops a deep learning-based system to classify movie reviews as positive or negative.The IMDB dataset of 50,000 labeled reviews was used. Data preprocessing included removing HTML tags, normalization, tokenization, stop-word removal, and stemming. The processed text was converted into numerical form using word embeddings to capture meaning.A Bidirectional LSTM model was used, which reads text in both directions to understand context better than traditional models. The model was trained with Adam optimizer and Binary Cross-Entropy loss, with early stopping to reduce overfitting.Results showed about 88.5% accuracy, proving the model’s effectiveness. This approach can be applied in recommendation systems, social media analysis, and feedback evaluation. Future work may include using advanced models like BERT and handling sarcasm and multilingual data.
