Output details
11 - Computer Science and Informatics
Robert Gordon University
Output title
Visualizing and Evaluating Complexity of Textual Case Bases
Type
E - Conference contribution
Name of conference/published proceedings
Advances in Case-Based Reasoning: Proceedings of the 9th European Conference, ECCBR 2008 (LNCS Volume 5239)
Volume number
5239
Issue number
-
First page of article
104
ISSN of proceedings
1611-3349
Year of publication
2008
URL
-
Number of additional authors
5
Additional information
This paper introduces a novel approach to visualization of textual case bases.
The method takes the case-term matrix and replacing its ones and zeros with light and dark squares. The columns and rows are then re-arranged according to their pairwise similarity, leading to an image with larger blocks of colour. This makes clusters of similar documents more visible and the overall pattern gives an idea of the complexity of the case base.
A complexity metric based on the visualization was a good predictor of classifier performance, with the expected result that lower accuracy was achieved for more complex case bases
Interdisciplinary
-
Cross-referral requested
-
Research group
None
Citation count
1
Proposed double-weighted
No
Double-weighted statement
-
Reserve for a double-weighted output
No
Non-English
No
English abstract
-
