Neural Network-Based Approach to Literary Selection for Grades 5-9

The selection of relevant literature for middle school students is crucial for enhancing their intellectual growth and fostering their interest in reading. This study presents a neural network-based natural language processing approach for optimizing literary selection. Texts from the “School Literature Corpus,” derived from Uzbek school literature textbooks, were preprocessed and represented as embeddings using the transformer model all-MiniLM-L6-v2. By applying cosine similarity, literary works were aligned with the intellectual and vocabulary levels of specific grades, enabling personalized and context-aware recommendations. The results demonstrate that this method yields more meaningful literature recommendations, enriching curricula and fostering critical thinking, literacy, and a lifelong love of reading. This approach provides educators and institutions with learning materials tailored to the intellectual potential of their students, supporting the development of knowledgeable, curious, and intellectually gifted learners. Furthermore, it ensures continuous improvement in educational quality and contributes to the advancement of society's future.