Digital Transformation of English Literary Studies: Computational Approaches to Narrative Analysis, Thematic Interpretation, and Literary Patterns

The digital revolution of literary studies has opened up new possibilities for combining computational tools and approaches with more traditional textual ones. This research explores the use of natural language processing (NLP), text mining, topic modelling, and machine learning to detect the patterns of narration, theme, and literary elements of selected English novels. A structured corpus of 1,000 consecutive text fragments of Jane Austen's Pride and Prejudice, Mary Shelley's Frankenstein, and Bram Stoker's Dracula has been analysed computationally. To analyse the textual characteristics and cross-novel differences, text preprocessing, descriptive analysis, lexical analysis, Term Frequency-Inverse Document Frequency (TF-IDF), Non-negative Matrix Factorization (NMF), semantic similarity analysis and K-means clustering were used. The analysis showed that there were differences in word count and distinctive lexical profile at the segment level among the three works. NMF revealed six interpretable thematic structures with good novel-specific thematic distributions and higher thematic heterogeneity in Dracula. The literary works were especially well clustered, with 98.9% of text segments clustered in the direction of their source novels and an Adjusted Rand Index of 0.971. The results show that the computational approach can reveal latent lexical, thematic and structural patterns in literary text. The study demonstrates that, with the support of systematic and reproducible textual evidence, computational analysis can complement the traditional literary interpretation, and it can be used to strengthen interdisciplinary linkage among English literary studies, artificial intelligence and intelligent information processing.