Project Update

1
Project Update
XML Document Visualization and Retrieval

• Matt Williams

2
Background

• Can we take advantage of this structure when
  searching for documents?

• XML vs Web Doc
• Added Structure

ltbookgt lttitlegtMy First XMLlt/titlegt ltprod
id"33-657 media"paper"gt lt/prodgt
ltchaptergtIntroduction to XML ltparagtWhat is
HTMLlt/paragt ltparagtWhat is XMLlt/paragt
lt/chaptergt ltchaptergtXML Syntax
ltparagtElements must have a closing taglt/paragt
ltparagtElements must be properly nestedlt/paragt
lt/chaptergt lt/bookgt
3
Information Retrieval

• Standard Information Retrieval (IR)
• tfidf
• tf frequency of a term in a doc
• Idf inverse document frequency
• Number of documents containing the term

4
Information Retrieval

• A fair bit of previous work on adding structure
  to IR queries.
• Examples
• XIRQL Fuhr and GroBjohann
• //book/chapterheading cw InfoVis
• XXL Theobald and Weikum
• Select Z From Index
• Where zoos.animal.cougar as Z
• But
• What if we are unsure of the structure?
• What if we have variability in the structure?

5
Information Retrieval

• My goal is to provide an interface to explore the
  XML collection with limited information
• Meta-Schema Information Element Index
• Visual Clustering Multidimensional Scaling
• Visual Queries Element Selection

6
Related Work

• Visual Information Seeking
• Homefinder / Periodic Table Algerg and
  Shneiderman

7
Related Work

• Galaxies
• Wise et al.
• Visual Web Retrieval
• Lighthouse - Leuski

8
Related Work

• ZUI Pad, Jazz, and Piccolo
• Ben Bederson
• SpaceTree
• Jesse Grosjean et al.
• TreeMaps ??
• Ben Shneiderman

9
Multidimensional Scaling

• Document Similarity
• Dimensionality Reduction From full dimensional
  distance measure ? 2 dimensional distance measure
• Problems Speed?

10
Test Environment

• eXist Open Source XML Native Database
• Wolfgang M. Meier
• http//exist-db.org/
• I am working on providing a front end to the
  Database that provides
• A Selectable Element Index
• Interactive Results That Dynamically Cluster and
  Zoom

11
Thus Far

• Lots of Learning!!
• XML Databases
• Multidimensional Scaling
• XML Queries
• XML Information Retrieval
• Zoomable Interfaces
• Treemaps
• Added basic GUI to eXist
• Added a Service to offer the element Index as
  part of the API