F. Denis. Pac learning from positive statistical queries. In Algorithmic Learning Theory (ALT), 9th International Conference, volume 1501 of LNAI, pages 112--126. Springer, 1998.  Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
PEBL: Positive Example Based Learning for Web Page Classification .. - Yu, Han (2002) (13 citations) (Correct)
....stream of research which uses unlabeled data in classification is termed learning from positive and unlabeled data. In 1998, F. Denis defined the PAC learning model for positive and unlabeled examples, and showed that kDNF (Disjunctive Normal Form) is learnable from positive and unlabeled examples [7]. Since then, some experimental attempts to learn using positive and unlabeled data have been tried using k DNF or decision trees [13, 5] However, those methods are not very useful for Web page classification problems because; 1) k DNF or decision trees are not very tolerant with high ....
....the one class SVM does not utilize the distribution of unlabeled data. Our approach is fundamentally different from previous approaches. The first stage of the M C algorithm (called the mapping stage) is based on 1 DNF, which was previously proven learnable from positive and unlabeled data [7]. The 2http: www.csie.nt u.edu.tw cjlin libsvm Figure 2: A graphical representation of a linear SVM in a two dimensional case. i.e. Only two features are considered. M is the distance from the separator to the support vectors in feature space. second stage of the M C algorithm (coiled the ....
[Article contains additional citation context not shown here]
F. Denis. Pac learning from positive statistical queries. In ALT, 1998.
Learning from Positive and Unlabeled Examples - Letouzey, Denis, Gilleron (2000) (5 citations) Self-citation (Denis) (Correct)
No context found.
F. Denis. PAC learning from positive statistical queries. In Michael M. Richter, Carl H. Smith, Rolf Wiehagen, and Thomas Zeugmann, editors, Proceedings of the 9th International Conference on Algorithmic Learning Theory (ALT-98), volume 1501 of LNAI, pages 112-126, Berlin, October 8{ 10 1998. Springer.
Positive and Unlabelled Examples Help Learning - De Comit, Denis, Gilleron.. Self-citation (Denis) (Correct)
....too: for a web page classi cation problem, unlabelled web pages can be inexpensively gathered, a set of web pages you are interested in is available in your bookmarks, labelled web pages are fairly expensive but a small set of hand labelled web pages can be designed. It has been proved in [Den98] that many concepts classes, namely those which are learnable from statistical queries, can be e ciently learned in a PAC framework using positive and unlabelled data only. But the price to pay is a considerable increase in the number of examples needed to achieve learning (although it remains of ....
F. Denis. Pac learning from positive statistical queries. In ALT 98, 9th International Conference on Algorithmic Learning Theory, volume 1501 of Lecture Notes in Arti cial Intelligence, pages 112126. Springer-Verlag, 1998.
Learning Regular Languages From Simple Positive Examples - Denis (1998) (6 citations) Self-citation (Denis) (Correct)
....and shown the existence of polynomially computable characteristic samples sufficient for identification from positive data. See also [Yok95] HKY98] Due to the free distribution and polynomial running time requirements, results are still weaker in PAC learning framework ( Nat91a] Shv90] [Den98]) Gold suggests that one of the reason why natural language learning is possible is that the learner is not provided with arbitrary examples. There are several ways to give substance to this idea: ffl the learner may asks queries. Angluin proved in [Ang87] that REG is exactly learnable within ....
....= L and Learn(Teach(L) L 0 , which is contradictory. Unfortunately, the situation is even getting worse in the PAC learning model. It can be easily shown that if a class is PAC learnable from positive data (as the class of k CNF) the output hypothesis must be included in the target concept ([Den98]) But as it is impossible from positive data to differentiate a negative example from an absent positive one, even very simple classes cannot be PAC learnable from positive data. See [Nat91a] Shv90] Den98] for a detailed study. See also [Sak92] for results on grammatical inference about ....
[Article contains additional citation context not shown here]
F. Denis. Pac learning from positive statistical queries. In ALT 98, 9th International Conference on Algorithmic Learning Theory, volume 1501 of Lecture Notes in Artificial Intelligence, pages 112--126. Springer-Verlag, 1998.
Journal of Machine Learning Research 7 (2006) 283--306.. - Paul Goldberg Pwg (Correct)
No context found.
F. Denis. Pac learning from positive statistical queries. In Algorithmic Learning Theory (ALT), 9th International Conference, volume 1501 of LNAI, pages 112--126. Springer, 1998.
Learning to Filter Junk E-Mail from Positive and Unlabeled.. - Schneider (2004) (Correct)
No context found.
F. Denis. PAC learning from positive statistical queries. Workshop on Algorithmic Learning Theory (ALT'98), 112--126, 1998.
PEBL: Web Page Classification without Negative Examples - Yu, Han, Chang (2004) (Correct)
No context found.
F. Denis, "PAC Learning from Positive Statistical Queries," Proc. 10th Int'l Conf. Algorithmic Learning Theory (ALT '99), pp. 112-126, 1998.
Building Text Classifiers Using Positive and Unlabeled Examples - Liu, Dai, Li, al. (2003) (4 citations) (Correct)
No context found.
.# Denis, F. PAC learning from positive statistical queries. ALT-98.
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