Sentiment Analysis of Literary Texts: A Study of Theme and Readers’ Preferences in Russian Short Stories from 1900 to 1930s
摘要
This article presents an experimental analysis of Russian short stories from the early twentieth century, focusing on the sentiment of literary pieces in the context of their themes and readers’ preferences toward these texts. The study utilized a corpus of 210 texts, spanning from 1900 to 1930, each manually thematically annotated by experts. Sentiment was assessed employing sentiment lexicons, specifically RuSentiLex and the Multilingual Sentiment Lexicon, across individual stories and stories grouped by theme. Additionally, an experiment aimed at identifying a correlation between the texts’ sentiment and readers’ preferences was conducted. Findings indicate a predominant use of words from negative lexicons in the Russian short stories of the period. An association between thematic tags and the sentiment of texts was discovered, showing that stories with positive vocabulary often correlate with themes like FUTURE, YOUTH, LEISURE, VICES, and FREEDOM. Conversely, stories with a mainly negative lexical presence commonly carry tags of DEATH, WAR, VIOLENCE, and, unexpectedly, CHILDREN. Furthermore, the study highlighted an absence of correlation between the sentiment portrayed in the texts and the reader preferences, suggesting a complex dynamic between narrative tone and reader reception.