Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Longitudinal Abuse and Sentiment Analysis of Hollywood Movie Dialogues using Language Models

Rohitash Chandra, Guoxiang Ren, Group-H

TL;DR

This study conducts a longitudinal analysis of Hollywood movie dialogues from 1950 to 2024 to quantify trends in abusive content and sentiment using fine-tuned language models. The authors curate a large subtitle corpus of Oscar-nominated and top-grossing films, classify them into four genres, and apply HateBERT for abuse detection and RoBERTa-based SenWave models for multi-label sentiment analysis, complemented by decadal n-gram analyses. Key findings reveal a rise in abusive language over time, with Oscar-nominated films increasingly surpassing blockbusters in recent decades, and persistent humour as a dominant positive sentiment across genres. The work offers insights into how cinematic language reflects and potentially shapes social norms, with implications for content moderation, ratings, and marketing, and suggests future multimodal extensions to incorporate visual and audio cues.

Abstract

Over the past decades, there has been an increase in the prevalence of abusive and violent content in Hollywood movies. In this study, we use language models to explore the longitudinal abuse and sentiment analysis of Hollywood Oscar and blockbuster movie dialogues from 1950 to 2024. We provide an analysis of subtitles for over a thousand movies, which are categorised into four genres. We employ fine-tuned language models to examine the trends and shifts in emotional and abusive content over the past seven decades. Findings reveal significant temporal changes in movie dialogues, which reflect broader social and cultural influences. Overall, the emotional tendencies in the films are diverse, and the detection of abusive content also exhibits significant fluctuations. The results show a gradual rise in abusive content in recent decades, reflecting social norms and regulatory policy changes. Genres such as thrillers still present a higher frequency of abusive content that emphasises the ongoing narrative role of violence and conflict. At the same time, underlying positive emotions such as humour and optimism remain prevalent in most of the movies. Furthermore, the gradual increase of abusive content in movie dialogues has been significant over the last two decades, where Oscar-nominated movies overtook the top ten blockbusters.

Longitudinal Abuse and Sentiment Analysis of Hollywood Movie Dialogues using Language Models

TL;DR

This study conducts a longitudinal analysis of Hollywood movie dialogues from 1950 to 2024 to quantify trends in abusive content and sentiment using fine-tuned language models. The authors curate a large subtitle corpus of Oscar-nominated and top-grossing films, classify them into four genres, and apply HateBERT for abuse detection and RoBERTa-based SenWave models for multi-label sentiment analysis, complemented by decadal n-gram analyses. Key findings reveal a rise in abusive language over time, with Oscar-nominated films increasingly surpassing blockbusters in recent decades, and persistent humour as a dominant positive sentiment across genres. The work offers insights into how cinematic language reflects and potentially shapes social norms, with implications for content moderation, ratings, and marketing, and suggests future multimodal extensions to incorporate visual and audio cues.

Abstract

Over the past decades, there has been an increase in the prevalence of abusive and violent content in Hollywood movies. In this study, we use language models to explore the longitudinal abuse and sentiment analysis of Hollywood Oscar and blockbuster movie dialogues from 1950 to 2024. We provide an analysis of subtitles for over a thousand movies, which are categorised into four genres. We employ fine-tuned language models to examine the trends and shifts in emotional and abusive content over the past seven decades. Findings reveal significant temporal changes in movie dialogues, which reflect broader social and cultural influences. Overall, the emotional tendencies in the films are diverse, and the detection of abusive content also exhibits significant fluctuations. The results show a gradual rise in abusive content in recent decades, reflecting social norms and regulatory policy changes. Genres such as thrillers still present a higher frequency of abusive content that emphasises the ongoing narrative role of violence and conflict. At the same time, underlying positive emotions such as humour and optimism remain prevalent in most of the movies. Furthermore, the gradual increase of abusive content in movie dialogues has been significant over the last two decades, where Oscar-nominated movies overtook the top ten blockbusters.
Paper Structure (20 sections, 17 figures, 8 tables)

This paper contains 20 sections, 17 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Cinema: impact on society
  4. Web series and movies
  5. Methodology
  6. Data
  7. N-gram analysis
  8. Model Selection and Fine-tuning Strategy
  9. Framework
  10. Technical Details
  11. Results
  12. Data analysis
  13. Sentiment analysis
  14. Abuse and Hate-Speech detection
  15. Case studies
  16. ...and 5 more sections

Figures (17)

  • Figure 1: Our sentiment analysis and abusive detection framework for selected Hollywood movies from 1950 to 2024.
  • Figure 2: Top 10 bigrams and trigrams for different decades showing longitudinal changes.
  • Figure 3: Heatmap about Sentiments Co-occurrence in movies during 75 years
  • Figure 4: Counts of different emotions for Oscar-nominated and top ten blockbuster movies from 1950-2024.
  • Figure 5: Emotion Counts for Different Categories
  • ...and 12 more figures