Abstract
Although Behavior-Knowledge Space (BKS) method does not need any assumptions in combining multiple experts, it should build theoretically exponential storage spaces for storing and managing jointly observed K decisions from K experts. That is, combining K experts needs a (K+1)st-order probability distribution. However, it is well known that the distribution becomes unmanageable in storing and estimating, even for a small K. In order to overcome such weakness, it would be attractive to decompose the distribution into a number of component distributions and to approximate the distribution with a product of the component distributions. One of such previous works is to apply a conditional independence assumption to the distribution. Another work is to approximate the distribution with a product of only first-order tree dependencies or second-order distributions as shown in [1]. In this paper, a dependency-based framework is proposed to optimally approximate a probability distribution with a product set of dth-order dependencies where 1≤d≤K, and to combine multiple experts based on the product set using the Bayesian formalism. This framework was experimented and evaluated with a standardized CEN-PARMI data base.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition |
| Publisher | IEEE |
| Pages | 124-129 |
| Number of pages | 6 |
| Volume | 2 |
| Publication status | Published - 1999 |
| Event | Proceedings of the 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'99) - Fort Collins, CO, USA Duration: 1999 Jun 23 → 1999 Jun 25 |
Other
| Other | Proceedings of the 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'99) |
|---|---|
| City | Fort Collins, CO, USA |
| Period | 99/6/23 → 99/6/25 |
Fingerprint
ASJC Scopus subject areas
- Computer Vision and Pattern Recognition
- Software
- Control and Systems Engineering
- Electrical and Electronic Engineering
Cite this
Dependency-based framework of combining multiple experts for the recognition of unconstrained handwritten numerals. / Kang, Hee Joong; Lee, Seong Whan.
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Vol. 2 IEEE, 1999. p. 124-129.Research output: Chapter in Book/Report/Conference proceeding › Chapter
}
TY - CHAP
T1 - Dependency-based framework of combining multiple experts for the recognition of unconstrained handwritten numerals
AU - Kang, Hee Joong
AU - Lee, Seong Whan
PY - 1999
Y1 - 1999
N2 - Although Behavior-Knowledge Space (BKS) method does not need any assumptions in combining multiple experts, it should build theoretically exponential storage spaces for storing and managing jointly observed K decisions from K experts. That is, combining K experts needs a (K+1)st-order probability distribution. However, it is well known that the distribution becomes unmanageable in storing and estimating, even for a small K. In order to overcome such weakness, it would be attractive to decompose the distribution into a number of component distributions and to approximate the distribution with a product of the component distributions. One of such previous works is to apply a conditional independence assumption to the distribution. Another work is to approximate the distribution with a product of only first-order tree dependencies or second-order distributions as shown in [1]. In this paper, a dependency-based framework is proposed to optimally approximate a probability distribution with a product set of dth-order dependencies where 1≤d≤K, and to combine multiple experts based on the product set using the Bayesian formalism. This framework was experimented and evaluated with a standardized CEN-PARMI data base.
AB - Although Behavior-Knowledge Space (BKS) method does not need any assumptions in combining multiple experts, it should build theoretically exponential storage spaces for storing and managing jointly observed K decisions from K experts. That is, combining K experts needs a (K+1)st-order probability distribution. However, it is well known that the distribution becomes unmanageable in storing and estimating, even for a small K. In order to overcome such weakness, it would be attractive to decompose the distribution into a number of component distributions and to approximate the distribution with a product of the component distributions. One of such previous works is to apply a conditional independence assumption to the distribution. Another work is to approximate the distribution with a product of only first-order tree dependencies or second-order distributions as shown in [1]. In this paper, a dependency-based framework is proposed to optimally approximate a probability distribution with a product set of dth-order dependencies where 1≤d≤K, and to combine multiple experts based on the product set using the Bayesian formalism. This framework was experimented and evaluated with a standardized CEN-PARMI data base.
UR - http://www.scopus.com/inward/record.url?scp=0032638722&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0032638722&partnerID=8YFLogxK
M3 - Chapter
VL - 2
SP - 124
EP - 129
BT - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
PB - IEEE
ER -