H. Lounis and G. Bisson, Evaluation of Learning Systems: An Artificial Data-Based Approach, In Proceedings of European Working Session on Learning, pp: 463-481, 1991. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Batch Learning of Disjoint Feature Intervals - Akkus (1996) (1 citation) (Correct)
....108 liver 129 148 267 114 607 musk 3,477 9,529 11,279 2,740 18,744 vehicle 586 4,787 818 2,012 14,441 wine 113 79 73 299 600 5.2. 3 Experiments with Artificial Datasets To cope with noisy and incomplete data is an important criteria for a learning system to be used in real world applications [40]. One important point for a learning system is presence of irrelevant features [9] Therefore, artificial datasets are important to study the effects of irrelevant features, noise in the domain, and missing feature values since artificial datasets allows to test the system in a more controlled ....
H. Lounis and G. Bisson, Evaluation of Learning Systems: An Artificial Data-Based Approach, In Proceedings of European Working Session on Learning, pp: 463-481, 1991.
Non-Incremental Classification Learning Algorithms Based On.. - Demiröz (Correct)
....on Artificial Datasets Real world domains might have irrelevant features, noisy instances, and unknown (missing) feature values. Learning even in the presence of irrelevant features is an important criteria for a learning system [10] as well as learning from noisy and or incomplete data [47]. Therefore, we generated artificial datasets from a real world dataset by adding irrelevant features, noise, and unknown values and empirically evaluated the VFI algorithms compared with other classifiers. We used the real world dataset Iris in our experiments and observed the change in the ....
H. Lounis and G. Bisson, Evaluation of Learning Systems: An Artificial Data--Based Approach, In Proceedings of European Working Session on Learning, 463--481, 1991.
Classification With Overlapping Feature Intervals - Koc (1995) (Correct)
....4.2. 2 Experiments with Artificial Datasets The success of a learning system is very dependent on the ability to cope with noisy and incomplete data, an adequate knowledge representation scheme, having a low learning and prediction complexities and the effectiveness of the learned knowledge [34]. Most of the real world datasets contain incomplete and inconsistent data, therefore the ability to handle the inconsistent and incomplete data is very important for learning algorithms. Many researchers tackled the problem of handling examples which contain CHAPTER 4. EVALUATION OF THE COFI ....
H. Lounis and G. Bisson, Evaluation of Learning Systems: An Artificial Data-Based Approach, In Proceedings of European Working Session on Learning, pp: 463-481, 1991.
Learning with Feature Partitions - Sirin, Güvenir (1994) (1 citation) (Correct)
....for the maintenance and evolution of expert systems. The success of a learning system is highly related to the ability to cope with noisy and incomplete data, an adequate knowledge representation scheme, having low learning and sample complexities, and the effectiveness of the learned knowledge [24]. Tradeoffs between learning and programming can also be examined in terms of their relative utilities. Computer time is now very cheap, whereas human labor is becoming increasingly expensive. This suggests that learning could have an immediate economic edge over manual methods of programming the ....
....generated data sets are presented. In the artificially generated data sets some of the domain variables (such as number of features, number of examples, amount of noise, unknown attribute values, irrelevant attributes) are changed to test the behavior of the system under different conditions [24]. In Section 2 the effects of the domain dependent parameters of the CFP are also investigated under different settings. Third section presents the performance of the CFP and GA CFP on real data sets, and comparisons with other similar systems. Parameters of the Genetic Algorithm used by GA CFP: ....
[Article contains additional citation context not shown here]
H. Lounis and G. Bisson. Evaluation of Learning Systems: An Artificial Data-Based Approach. In Proceedings of European Working Session on Learning, pages 463--481, 1991.
Non-Incremental Classification Learning Algorithms Based On.. - Demiröz (1997) (Correct)
....on Artificial Datasets Real world domains might have irrelevant features, noisy instances, and unknown (missing) feature values. Learning even in the presence of irrelevant features is an important criteria for a learning system [10] as well as learning from noisy and or incomplete data [47]. Therefore, we generated artificial datasets from a real world dataset by adding irrelevant features, noise, and unknown values and empirically evaluated the VFI algorithms compared with CHAPTER 5. EVALUATION OF THE VFI ALGORITHMS 96 other classifiers. We used the real world dataset Iris in our ....
H. Lounis and G. Bisson, Evaluation of Learning Systems: An Artificial Data--Based Approach, In Proceedings of European Working Session on Learning, 463--481, 1991.
A Comprehensive Case Study: An Examination of Machine Learning.. - Zarndt (1995) (5 citations) (Correct)
No context found.
H. Lounis and G. Bisson (1991). Evaluation of Learning Systems: An Artificial Data-Based Approach. Lecture Notes in Artificial Intelligence Vol. 482 J. Siekmann editor. Springer-Verlag, Berlin. 463-481.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC