243 search results - page 40 / 49 » Interpretation of Trained Neural Networks by Rule Extraction |
MCS
2002
Springer
13 years 10 months ago
2002
Springer
The increasing amount and complexity of data in toxicity prediction calls for new approaches based on hybrid intelligent methods for mining the data. This focus is required even mo...
ICCV
2003
IEEE
15 years 21 days ago
2003
IEEE
In this paper we show that efficient object recognition can be obtained by combining informative features with linear classification. The results demonstrate the superiority of in...
ICIP
2003
IEEE
15 years 11 days ago
2003
IEEE
In this paper we present a system for classifying various human actions in compressed domain video framework. We introduce the notion of quantifying the motion involved, through w...
VLDB
1989
ACM
14 years 2 months ago
1989
ACM
This paper presents a parallel algorithm for recursive query processing and shows how it can be efficiently implemented in a local computer network. The algorithm relies on an int...
IJPRAI
2002
13 years 10 months ago
2002
Decision-tree algorithms are known to be unstable: small variations in the training set can result in different trees and different predictions for the same validation examples. B...