Zobrazeno 1 - 5
of 5
pro vyhledávání: '"Iván Piza-Dávila"'
Un algoritmo de clasificación incremental basado en los k vecinos más similares para datos mezclados
Autor:
Guillermo Sánchez-Díaz, Uriel E. Escobar-Franco, Luis R. Morales-Manilla, Iván Piza-Dávila, Carlos Aguirre-Salado, Anilu Franco-Arcega
Publikováno v:
Revista Facultad de Ingeniería Universidad de Antioquia, Iss 67, Pp 19-30 (2013)
En este trabajo, se presenta un algoritmo de clasificación incremental basado en los k vecinos más similares, el cual permite trabajar con datos mezclados y funciones de semejanza que no necesariamente son distancias. El algoritmo presentado es ade
Externí odkaz:
https://doaj.org/article/4bade34c022d4400b9bc57ab49fe2673
Autor:
Guillermo Sánchez-Díaz, Uriel E. Escobar-Franco, Luis R. Morales-Manilla, Iván Piza-Dávila, Carlos Aguirre-Salado, Anilu Franco-Arcega
Publikováno v:
Revista Facultad de Ingeniería Universidad de Antioquia, Iss 67 (2013)
This paper presents an incremental k-most similar neighbor classifier, for mixed data and similarity functions that are not necessarily distances. The algorithm presented is suitable for processing large data sets, because it only stores in main memo
Externí odkaz:
https://doaj.org/article/de54615b91ff488cb0e5a7abf9884a57
Publikováno v:
IEEE Access, Vol 8, Pp 56312-56320 (2020)
In pattern recognition, the elimination of unnecessary and/or redundant attributes is known as feature selection. Irreducible testors have been used to perform this task. An objective of the Minimum Description Length Principle (MDL) applied to featu
Externí odkaz:
https://doaj.org/article/891763a20683423ba838a96c5f29841e
Publikováno v:
Applied Sciences, Vol 9, Iss 20, p 4381 (2019)
Anomaly-based intrusion detection systems use profiles to characterize expected behavior of network users. Most of these systems characterize the entire network traffic within a single profile. This work proposes a user-level anomaly-based intrusion
Externí odkaz:
https://doaj.org/article/390966030f3941179abf8183872f16b7
Publikováno v:
International Journal of Computational Intelligence Systems, Vol 5, Iss 6 (2012)
In this paper, we introduce a fast implementation of the CT EXT algorithm for testor property identification, that is based on an accumulative binary tuple. The fast implementation of the CT EXT algorithm (one of the fastest algorithms reported), is
Externí odkaz:
https://doaj.org/article/089c94d4207644c497188956159c318e