Conclusion

Our exploration revealed that the genre of a book is not as straightforward as its label might suggest. Books often blend various themes and styles, making them more diverse than their main genre label indicates. This finding challenges the conventional method of categorizing books and urges us to consider the multiple dimensions within a single work. 

While the ECCO-BERT-Seq classifier excelled in identifying genres with distinct textual structures, it also illuminated the challenges in classifying more ambiguous genres. This points to the need for more balanced training data and a deeper understanding of the subtle cues that define different genres. Despite these challenges, our study has significantly enhanced our understanding of literary genres. It has provided new insights into the nature of 18th-century literature, demonstrating the complex interplay of genres within individual books.

Looking forward, as we think of future literary genre analysis, it is clear that there are many opportunities for enhancing and expanding our methodologies. Training the classifier with a broader and more diverse dataset stands as a crucial step towards improving its efficacy and sensitivity to the nuances of different genres. Moreover, delving deeper into the texts, perhaps moving beyond chapter-level analysis to examine smaller textual segments, could provide more detailed insights into the subtle shifts and patterns in genre usage. A comprehensive corpus analysis would also be instrumental in advancing our understanding of genre distinctions. Such an analysis could explore the specific linguistic features, thematic structures, and stylistic elements that define and differentiate genres.

 

• < Previous section: Analysis | Next section: Division of Labour and Reflection on Learning >