Impact Factor

7.883 Call For Paper

Share on

Author(s):

Thalari Murali | S.Om Prakash Reddy | SK. Sai Shakeer | VV. Shabeer
Keywords:

Bi-LSTM, GloVe Embeddings, Natural Language Processing, Sentiment Analysis, Spoiler Detection, Text Classification
Abstract:
The Rapid Proliferation Of Online Movie Reviews Has Created A Pressing Need For Automated Systems Capable Of Distinguishing Spoiler Content From Non-spoiler Opinions. This Paper Presents The Smart Review Analysis System (SRAS), An Intelligent Spoiler Detection Framework Built Upon A Bidirectional Long Short-Term Memory (Bi-LSTM) Neural Network Augmented With Pre-trained GloVe Word Embeddings. The Proposed System Processes IMDB User Reviews And Classifies Them As Spoiler Or Non-spoiler With High Accuracy. Extensive Preprocessing, Tokenization, And Sequence Padding Are Applied To The Textual Data Prior To Model Training. The Architecture Employs Stacked Bi-LSTM Layers, Spatial Dropout For Regularization, And A Sigmoid Output Layer For Binary Classification. Experimental Results On The IMDB Spoiler Dataset Demonstrate That SRAS Achieves Competitive Classification Performance, Validated Through Accuracy, Precision, Recall, And F1-score Metrics. The System Provides A Practical And Scalable Solution For Real-time Spoiler Filtering In Movie Review Platforms, Enhancing User Experience And Content Discovery.