Presented By O’Reilly and Cloudera
Make Data Work
September 11, 2018: Training & Tutorials
September 12–13, 2018: Keynotes & Sessions
New York, NY

From emotion analysis and topic extraction to narrative modeling

Andreea Kremm (Netex Group), Mohammed Ibraaz Syed (UCLA)
5:25pm–6:05pm Wednesday, 09/12/2018
Data science and machine learning
Location: 1A 08 Level: Beginner
Secondary topics:  Text and Language processing and analysis
Average rating: ****.
(4.00, 2 ratings)

Who is this presentation for?

  • Social and behavioral science enthusiasts and those working with natural language processing and understanding

Prerequisite knowledge

  • A basic understanding of topic modeling and natural language processing and understanding (useful but not required)

What you'll learn

  • Understand narrative economics and narrative modeling (and how the two are related)
  • Explore current work in narrative modeling

Description

Introduced formally by the renowned economist Robert Shiller in January 2017, narrative economics studies the impact of popular narratives and stories on economic fluctuations in the context of human interests and emotions. Professor Shiller is one of few economists to have predicted both the Great Recession of 2007–2009 and the dot-com crash. The term “narrative” refers to stories or ideas, particularly those of human interest and emotion that are spread through human communication. The human brain has been found to be highly attuned to narratives, whether factual or not, to justify ongoing decisions, actions, and even prejudices. Stories motivate and connect activities to deeply felt values and needs.

Narrative modeling is determining the narrative from human communication. Simply put, a narrative model summarizes and categorizes the narrative and associated emotion parameters. Narrative modeling is more challenging compared to sentiment analysis—the determination of positive, negative, or neutral sentiment in a piece of text. Narrative modeling involves emotion analysis, topic modeling, ER (entity-relation) extraction, and subject modeling. Andreea Kremm and Mohammed Ibraaz Syed describe the use of emotion analysis, entity relationship extraction, and topic modeling in modeling narratives from written human communication and share some of the work they’ve done in these areas.

Topics include:

  • Emotion analysis: Emotion extraction is a higher order operation compared to sentiment analysis. Extracting emotions from written text is a challenging area, especially for short text segments. Naveed and Rashed outline methods of determining emotions.
  • ER extraction: For modeling narrative, ER extraction is a key process to determine entities in text and their relationship in terms verbs. Entity extraction together with emotion analysis is the key to narrative modeling.
  • Topic extraction: A narrative needs to be summarized as well for it be useful in bigger schema of the human communication. Work involves adopting existing topic modeling methods for topic (of the narrative) extraction.
  • Subject modeling: One person may interpret a narrative differently from what another interpret it based upon their knowledge, thoughts, and prejudices. The subject model of a population determines the virality of a narrative.
Photo of Andreea Kremm

Andreea Kremm

Netex Group

Andreea Kremm is the founder of Netex Group, an international business service provider with over 700 employees in seven countries. Andreea has over 20 years of experience successfully implementing solutions for international online businesses, drawing on her expertise in computer science and psychology. She is currently focused on leveraging the power of AI and neural networks for economic applications through narrative economics, a relatively new research area that perfectly combines her two fields of expertise. Andreea holds a master’s degree in psychology from the University of Roehampton in London.

Photo of Mohammed Ibraaz Syed

Mohammed Ibraaz Syed

UCLA

Mohammed Ibraaz Syed recently completed his master’s degree in applied economics at UCLA, where he focused on utilizing data science and machine learning techniques to solve economic problems. One of his primary research interests is applying artificial intelligence (AI) algorithms to extract narratives from a corpus of text. Previously, Ibraaz worked at the World Bank, providing analysis of the bank’s existing work and developing databases that have been used to draw inferences and implications for improving the bank’s activities. Ibraaz holds a BA in economics and a BSc in mathematics from the University of Maryland, College Park.