Visual Data Mining: An Exploratory Approach to Analyzing Temporal Patterns of Eye Movements

Infant eye movements are an important behavioral resource to understand early human development and learning. But the complexity and amount of gaze data recorded from state‐of‐the‐art eye‐tracking systems also pose a challenge: how does one make sense of such dense data? Toward this goal, this artic...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Infancy Jg. 17; H. 1; S. 33 - 60
Hauptverfasser: Yu, Chen, Yurovsky, Daniel, Xu, Tian(Linger)
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Oxford, UK Blackwell Publishing Ltd 01.01.2012
Wiley-Blackwell
Wiley Subscription Services, Inc
Schlagworte:
Babies
Child development
Classification
Comparative Analysis
Data Analysis
Data mining
Eye Movements
Group Behavior
Individual Differences
Infants
Learning
Profiles
Reaction Time
Time
Visual Acuity
Visual Aids
Visualization
ISSN:1525-0008, 1532-7078, 1532-7078
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Infant eye movements are an important behavioral resource to understand early human development and learning. But the complexity and amount of gaze data recorded from state‐of‐the‐art eye‐tracking systems also pose a challenge: how does one make sense of such dense data? Toward this goal, this article describes an interactive approach based on integrating top‐down domain knowledge with bottom‐up information visualization and visual data mining. The key idea behind this method is to leverage the computational power of the human visual system. Thus, we propose an approach in which scientists iteratively examine and identify underlying patterns through data visualization and link those discovered patterns with top‐down knowledge/hypotheses. Combining bottom‐up data visualization with top‐down human theoretical knowledge through visual data mining is an effective and efficient way to make discoveries from gaze data. We first provide an overview of the underlying principles of this new approach of human‐in‐the‐loop knowledge discovery and then show several examples illustrating how this interactive exploratory approach can lead to new findings.
Bibliographie:ark:/67375/WNG-57TLJ387-S
istex:4F9B459DFE5BA3757B214D2554813C221C6D2DBB
ArticleID:INFA95
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 14
ObjectType-Article-1
ObjectType-Feature-2
content type line 23
ISSN:1525-0008
1532-7078
1532-7078
DOI:10.1111/j.1532-7078.2011.00095.x