The rapid expansion of digital media libraries has significantly increased the demand for intelligent
and personalized recommendation systems in modern streaming environments. This study proposes
a movie recommendation framework that incorporates Natural Language Processing (NLP)
techniques to improve the precision and relevance of suggested content. Unlike traditional
approaches that primarily depend on collaborative filtering or basic content-based strategies, which
often face challenges such as cold-start issues and shallow contextual interpretation, the proposed
method emphasizes deep textual analysis.
The system leverages NLP methodologies including data preprocessing, term weighting using TF
IDF, and similarity computation to extract meaningful insights from textual information such as
movie plots, genres, and user-generated reviews. Furthermore, advanced techniques such as word
embeddings and transformer-based architectures are considered to capture richer semantic
relationships within the data.
The evaluation of the proposed model indicates a noticeable improvement in recommendation
quality, particularly in terms of relevance and diversity, when compared to conventional methods.
The findings suggest that integrating NLP techniques into recommendation systems can enhance
user experience while maintaining scalability and interpretability, making the approach suitable for
real-world streaming applications.
