A boundary-fixed negative selection algorithm with online adaptive learning under small samples for anomaly detection

Liu, Dong; Liu, Shulin; Zhang, Hongli

doi:10.1016/j.engappai.2015.12.014

Cited by 26 publications

(11 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To determine the performance and possible advantage of ASTC-RNSA, we carried out experiments on 3 synthetic datasets [9], [19], [22] and 3 UCI datasets (Skin Segmentation Dataset, Iris Segmentation Dataset, Pima Indians Diabetes dataset) [17], [20], [21]. The 3 synthetic datasets (Ring, Cross, Pentagram) proposed by the intelligence security laboratory of Memphis University were introduced in this section and had been imported by many researches to verify the performance of NSA.…”

Section: Experiments and Discussionmentioning

confidence: 99%

“…Xiao et al [30] optimized the sequence of generated detectors with an immune mechanism, and hierarchically generated detectors from long to short distance to decrease the number of detectors. In 2016, Li et al [17] proposed a boundary-fixed negative selection algorithm based on online adaptive learning mechanism using small sample set for anomaly detection. It divides the feature space evenly into a large number of hypercubes and checks the relationship between every selfantigens and hypercube.…”

Section: Introductionmentioning

confidence: 99%

“…In essence, most of the variations of RNSA have focused on non-self-space coverage [11], [23], optimized detector generation [10], [13], [22], and minimized detector overlap [17]. However, the ''generate-discard'' mode has been the only detector generation model since the original NSA was proposed in 1994.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

An Antigen Space Triangulation Coverage Based Real-Value Negative Selection Algorithm

Zhang

Wen

et al. 2019

IEEE Access

View full text Add to dashboard Cite

Negative selection algorithm (NSA) is an important detector generation algorithm of artificial immune system (AIS). Traditional NSAs randomly generate detectors in the whole antigen space without considering the distribution of self/non-self antigens, therefore it is difficult for detectors to cover the whole antigen space evenly and a seriously time-consuming self-tolerance process is required. Aiming at the problem, we proposed a novel NSA based on antigen space triangulation coverage (ASTC) called ASTC-RNSA. In order to avoid the randomness in traditional NSAs, the proposed algorithm employed Delaunay Triangulation method from computational geometry to divide the self space into simplicial cells, which are utilized to determine the detector positions. Then, the overlaps between simplicial cells and self-antigens are removed to form a set of triangulation coverage areas. Finally, immune detectors which only covered non-self antigens space are generated in each triangulation coverage area. After Delaunay triangulation of antigen space, ASTC-RNSA can directly determine sui detector positions and avoid the time-consuming self-tolerance process of traditional NSAs. The analysis result shows that the time complexity of ASTC-RNSA is reduced from traditional exponential level to logarithmic level. The experimental results on several UCI datasets and artificial datasets show that the proposed algorithm can achieve more than 10 times the detector generation efficiency while maintaining similar detection performance with the most widely used representative typical algorithm, V-Detector.

show abstract

Section: Experiments and Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An Antigen Space Triangulation Coverage Based Real-Value Negative Selection Algorithm

Zhang

Wen

et al. 2019

IEEE Access

View full text Add to dashboard Cite

show abstract

“…One-class Classification (OCC) has been widely used for outlier, novelty, fault, and intrusion detection [1,2,3,4] by researchers from different disciplines. In OCC problems, samples of the class of interest (i.e., positive samples) are available while negative samples are very rare or costly to collect [5,6,7,8,9], thus making the application of multi-class models problematic. Various one-class classifiers [1,2] have been proposed based on the regression model, the clustering model etc.…”

Section: Introductionmentioning

confidence: 99%

“…Transformed output X h for the input X h−1 is computed to pass as the input to the next layer in the hierarchy. 7:…”

mentioning

confidence: 99%

Minimum Variance-Embedded Multi-layer Kernel Ridge Regression for One-class Classification

Gautam

Tiwari

Iosifidis

2018

2018 IEEE Symposium Series on Computational Intelligence (SSCI)

View full text Add to dashboard Cite

In this paper, a multi-layer architecture (in a hierarchical fashion) by stacking various Kernel Ridge Regression (KRR) based Auto-Encoder for one-class classification is proposed and is referred as MKOC. MKOC has many layers of Auto-Encoders to project the input features into new feature space and the last layer is regression based one class classifier. The Auto-Encoders use an unsupervised approach of learning and the final layer uses semi-supervised (trained by only positive samples) approach of learning. The proposed MKOC is experimentally evaluated on 15 publicly available benchmark datasets. Experimental results verify the effectiveness of the proposed approach over 11 existing state-of-the-art kernel-based one-class classifiers. Friedman test is also performed to verify the statistical significance of the claim of the superiority of the proposed one-class classifiers over the existing state-of-the-art methods. description [5], self-organizing map data description [5], Auto-Encoder data descriptor [11] etc. Whereas, the kernelbased one-class classifier approaches are support vector data description [12], one-class support vector machine[13], kernel principal component analysis based data description [14] etc. However, kernel-based methods have been shown to outperform non-kernel-based methods in the literature [1,5]. Despite this fact, these kernel-based methods involve the solution of a quadratic optimization problem, which is computationally expensive. Apart of these kernel-based methods, KRR-based models [15] optimize the problem rapidly in a non-iterative way by solving a linear systems.Therefore, KRR-based models [15,16,17,18,19] have received quite attention by researchers for solving various types of problems viz., regression, binary, multi-class etc. In recent years, various KRR-based one-class classifiers have been developed and exhibited better performance compared to various state-of-the-art one-class classifiers. Overall, the KRR-based one-class classifiers can be divided into two types, namely, (i) without Graph-Embedding (ii) with Graph-Embedding. For 'without Graph-Embedding', two types of architectures have been explored for OCC. One is KRR-based single output node architecture [20], and other is KRR-based Auto-Encoder architecture [21]. For 'with Graph-Embedding', Iosifidis et al.[22] presented local and global structure. Different types of Laplacian Graphs are employed for local (i.e., Local Linear Embedding, Laplacian Eigenmaps etc.) and global (linear discriminant analysis and clustering-based discriminant analysis etc.) Graph-embedding. Later, global variance-based Graph-Embedding has been extended in order to exploit class variance and sub-class variance information for face verification task by Mygdalis et al.[23]. All the above-mentioned KRR-based one-class classifiers employ a single-layered architectures.Over the last decade, stacked Auto-encoder based multi-layer architectures have received quite attention by researchers for multi-class or binary class classification tasks [24,...

show abstract

Self-adaptive mobile web service discovery approach based on modified negative selection algorithm

Garba

Mohamad

Saadon

2021

Neural Comput & Applic

View full text Add to dashboard Cite

A boundary-fixed negative selection algorithm with online adaptive learning under small samples for anomaly detection

Cited by 26 publications

References 28 publications

An Antigen Space Triangulation Coverage Based Real-Value Negative Selection Algorithm

An Antigen Space Triangulation Coverage Based Real-Value Negative Selection Algorithm

Minimum Variance-Embedded Multi-layer Kernel Ridge Regression for One-class Classification

Self-adaptive mobile web service discovery approach based on modified negative selection algorithm

Contact Info

Product

Resources

About