Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Three Stage Narrative Analysis; Plot-Sentiment Breakdown, Structure Learning and Concept Detection

Taimur Khan, Ramoza Ahsan, Mohib Hameed

TL;DR

The paper tackles automated narrative understanding by analyzing sentiment arcs, structure, and concepts in movie scripts. It presents a three-stage framework comprising Plot-Sentiment Breakdown with a custom NRC-VAD–LabMT lexicon, Structure Learning for segment classification, and Concept Detection for high/low concepts. Experiments on 1,000 scripts using Ward's hierarchical clustering reveal universal emotional arcs and demonstrate cross-genre similarities, suggesting practical value for publishers, writers, and researchers. Limitations include lexicon-based sentiment context and partial concept detection; future work envisions contextual embeddings, hybrid clustering with SOMs, and multilingual expansion to enhance robustness and applicability.

Abstract

Story understanding and analysis have long been challenging areas within Natural Language Understanding. Automated narrative analysis requires deep computational semantic representations along with syntactic processing. Moreover, the large volume of narrative data demands automated semantic analysis and computational learning rather than manual analytical approaches. In this paper, we propose a framework that analyzes the sentiment arcs of movie scripts and performs extended analysis related to the context of the characters involved. The framework enables the extraction of high-level and low-level concepts conveyed through the narrative. Using dictionary-based sentiment analysis, our approach applies a custom lexicon built with the LabMTsimple storylab module. The custom lexicon is based on the Valence, Arousal, and Dominance scores from the NRC-VAD dataset. Furthermore, the framework advances the analysis by clustering similar sentiment plots using Wards hierarchical clustering technique. Experimental evaluation on a movie dataset shows that the resulting analysis is helpful to consumers and readers when selecting a narrative or story.

Three Stage Narrative Analysis; Plot-Sentiment Breakdown, Structure Learning and Concept Detection

TL;DR

The paper tackles automated narrative understanding by analyzing sentiment arcs, structure, and concepts in movie scripts. It presents a three-stage framework comprising Plot-Sentiment Breakdown with a custom NRC-VAD–LabMT lexicon, Structure Learning for segment classification, and Concept Detection for high/low concepts. Experiments on 1,000 scripts using Ward's hierarchical clustering reveal universal emotional arcs and demonstrate cross-genre similarities, suggesting practical value for publishers, writers, and researchers. Limitations include lexicon-based sentiment context and partial concept detection; future work envisions contextual embeddings, hybrid clustering with SOMs, and multilingual expansion to enhance robustness and applicability.

Abstract

Story understanding and analysis have long been challenging areas within Natural Language Understanding. Automated narrative analysis requires deep computational semantic representations along with syntactic processing. Moreover, the large volume of narrative data demands automated semantic analysis and computational learning rather than manual analytical approaches. In this paper, we propose a framework that analyzes the sentiment arcs of movie scripts and performs extended analysis related to the context of the characters involved. The framework enables the extraction of high-level and low-level concepts conveyed through the narrative. Using dictionary-based sentiment analysis, our approach applies a custom lexicon built with the LabMTsimple storylab module. The custom lexicon is based on the Valence, Arousal, and Dominance scores from the NRC-VAD dataset. Furthermore, the framework advances the analysis by clustering similar sentiment plots using Wards hierarchical clustering technique. Experimental evaluation on a movie dataset shows that the resulting analysis is helpful to consumers and readers when selecting a narrative or story.

Paper Structure

This paper contains 20 sections, 10 figures, 1 table.

Table of Contents

  1. Introduction
  2. Motivation
  3. Background
  4. Problem Statement
  5. Related work
  6. Segmentation
  7. Sentiment Analysis
  8. Story Arcs
  9. Structure Analysis and Text Classification
  10. Semantics and Concept Extraction
  11. Proposed Work
  12. Custom Lexicon dictionary for LabMT emotion rating
  13. Segmentation of Script and Text Frequency vector
  14. Sentiment Analysis using LabMT
  15. Experimental Evaluation
  16. ...and 5 more sections

Figures (10)

  • Figure 1: Problem Breakdown: 3 Tier Narrative Analyzer; Plot-Sentiment Breakdown, Structure Learning and Concept Detection
  • Figure 2: Plot-Sentiment Breakdown for determining the Story Arc
  • Figure 3: Structure Learning and Text Classification: training procedure
  • Figure 4: Structure Learning and Text Classification: testing procedure
  • Figure 5: Dendrogram showing hierarchical clustering of 1,000 movie scripts using Ward’s linkage. The vertical axis represents the distance (variance) between sentiment trajectories, and the horizontal axis groups scripts with similar emotional arcs. Three primary clusters were identified, indicating recurring emotional patterns across different film genres.
  • ...and 5 more figures