Decision region connectivity analysis: A method for analyzing high-dimensional classifiers

Machine Learning

Volumn 48, Issue 1-3, 2002, Pages 321-351

  Melnik, Ofer ^a  

^a RUTGERS UNIVERSITY   (United States)

Author keywords

Classification; Concave; Convex; Decision boundary; Decision region; DRCA; High dimensional visualization; Minimum volume ellipsoid; PCA; Rule extraction

Indexed keywords

COMPUTATIONAL COMPLEXITY; MATHEMATICAL MODELS; NEURAL NETWORKS; TREES (MATHEMATICS);

DECISION REGION CONNECTIVITY ANALYSIS; HIGH-DIMENSIONAL VISUALIZATION; MINIMUM VOLUME ELLIPSOID; RULE EXTRACTION;

DECISION THEORY;

EID: 0036643059     PISSN: 08856125     EISSN: None     Source Type: Journal    
DOI: 10.1023/A:1013968124284     Document Type: Article

References (20)

1. Cascading classifiers
2. Covering a set of points by two axis-parallel boxes
3. UCI repository of machine learning databases
4. The maximum clique problem
5. Fast interactive 3-d graph visualization
6. Making large-scale svm learning practical
7. Statlolg: Comparison of classification algorithms on large real-world problems
8. An algorithmic theory of numbers, graphs and convexity
9. Theory and scope of exact representation extraction from feed-forward networks
10. Non deterministic polynomial optimization problems and their approximations
11. Introduction to positive-breakdown methods
12. Set estimation via ellipsoidal approximation
13. New algorithms for constructing optimal circumscribed and inscribed ellipsoids
14. Vehicle recognition using rule based methods
15. Separability is a learner's best friend
16. Estimation of correlation coefficients by ellipsoidal timming
17. Approximating the convex hull polyhedra with axis-parallel boxes