Self-adaptive negative selection using local outlier factor

Date

2012-09-05

Author

Ataser, Zafer
Alpaslan, Ferda Nur

Metadata

Show full item record

Item Usage Stats

123
views

0
downloads

Negative selection algorithm (NSA) classifies a given data either as normal (self) or anomalous (non-self). To make this classification, it is trained using normal (self) samples. NSA generates detectors to cover the complementary space of self in training phase. The classification of NSAs is mainly specified by two issues, self space determination and detectors coverage. The boundary of self is ambiguous so NSAs use self samples to calculate a space close to the self space. The other issue is the detectors coverage which should maximize non-self space coverage and minimize self space coverage. This paper introduces a novel NSA and this NSA proposes k-nearest neighbor and local outlier factor to determine self space for a given self samples. Beside these, it specifies the detectors coverage using Monte Carlo Integration. The experimental evaluations show that the novel NSA generates comparable and reasonable results.

Subject Keywords

Nearest neighbor search, Statistics

URI

https://hdl.handle.net/11511/88080

DOI

https://doi.org/10.1007/978-1-4471-4594-3_17

Conference Name

27th International Symposium on Computer and Information Sciences, ISCIS 2012, (3 - 05 Eylül 2012)

Collections

Department of Computer Engineering, Conference / Seminar

Suggestions

OpenMETU
Core

Similarity search in protein sequence databases using metric access methods Cetintas, Ahmet; Sacan, Ahmet; Toroslu, İsmail Hakkı (2013-09-13) The rapid increase in the size of biological sequence data owing to the advancements in high-throughput sequencing techniques, and the increased complexity of hypothesis-driven exploration of this data requiring massive number of similarity queries call for new approaches for managing sequence databases and analysis of this information. The metric space representation for sequences is suitable for similarity search and provides several sophisticated metric-indexing techniques. In this work, we provide a tho...
Robust estimation and hypothesis testing under short-tailedness and inliers Akkaya, Ayşen (Springer Science and Business Media LLC, 2005-06-01) Estimation and hypothesis testing based on normal samples censored in the middle are developed and shown to be remarkably efficient and robust to symmetric short-tailed distributions and to inliers in a sample. This negates the perception that sample mean and variance are the best robust estimators in such situations (Tiku, 1980; Dunnett, 1982).
Adaptive estimation and hypothesis testing methods Dönmez, Ayça; Tiku, Moti Lal; Department of Statistics (2010) For statistical estimation of population parameters, Fisher’s maximum likelihood estimators (MLEs) are commonly used. They are consistent, unbiased and efficient, at any rate for large n. In most situations, however, MLEs are elusive because of computational difficulties. To alleviate these difficulties, Tiku’s modified maximum likelihood estimators (MMLEs) are used. They are explicit functions of sample observations and easy to compute. They are asymptotically equivalent to MLEs and, for small n, are equal...
Quantum Search in Sets with Prior Knowledge Çalıkyılmaz, Umut; Turgut, Sadi; Department of Physics (2021-10-7) Quantum search algorithm revolutionized the field by reducing the complexity significantly for the search problem. However, by not being able to decrease the complexity to logarithmic scales, this algorithm still needs a significant amount of time to solve the search problem for large sets. It is proven that the order of the time required to solve this problem cannot be reduced further but, making an improvement by some constant factor is still possible. This aim has been pursued by some scientists in the p...
Non-normal bivariate distributions: estimation and hypothesis testing Qunsiyeh, Sahar Botros; Tiku, Moti Lal; Department of Statistics (2007) When using data for estimating the parameters in a bivariate distribution, the tradition is to assume that data comes from a bivariate normal distribution. If the distribution is not bivariate normal, which often is the case, the maximum likelihood (ML) estimators are intractable and the least square (LS) estimators are inefficient. Here, we consider two independent sets of bivariate data which come from non-normal populations. We consider two distinctive distributions: the marginal and the conditional dist...

Citation Formats

Z. Ataser and F. N. Alpaslan, “Self-adaptive negative selection using local outlier factor,” presented at the 27th International Symposium on Computer and Information Sciences, ISCIS 2012, (3 - 05 Eylül 2012), 2012, Accessed: 00, 2021. [Online]. Available: https://hdl.handle.net/11511/88080.