A feature selection method for classification within functional genomics experiments based on the proportional overlapping score

Mahmoud, Osama; Harrison, Andrew; Perperoglou, Aris; Gul, Asma; Khan, Zardad; Metodiev, Metodi V.; Lausen, Berthold

doi:10.1186/1471-2105-15-274

Cited by 43 publications

(64 citation statements)

References 38 publications

(71 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The method is implemented in an R package OTE. The proposed method could better be used, in conjunction with some feature selection method, (Mahmoud et al (2014a(Mahmoud et al ( , 2014b, for example) in high dimensional settings.…”

Section: Resultsmentioning

confidence: 99%

An Ensemble of Optimal Trees for Class Membership Probability Estimation

Khan

Gul

Mahmoud

et al. 2016

Analysis of Large and Complex Data

Self Cite

View full text Add to dashboard Cite

General rightsThis document is made available in accordance with publisher policies. Please cite only the published version using the reference above. Abstract. Machine learning methods can be used for estimating the class membership probability of an observation. We propose an ensemble of optimal trees in terms of their predictive performance. This ensemble is formed by selecting the best trees from a large initial set of trees grown by random forest. A proportion of trees is selected on the basis of their individual predictive performance on out-of-bag observations. The selected trees are further assessed for their collective performance on an independent training data set. This is done by adding the trees one by one starting from the highest predictive tree. A tree is selected for the final ensemble if it increases the predictive performance of the previously combined trees. The proposed method is compared with probability estimation tree, random forest and node harvest on a number of bench mark problems using Brier score as a performance measure. In addition to reducing the number of trees in the ensemble, our method gives better results in most of the cases. The results are supported by a simulation study.

show abstract

Section: Resultsmentioning

confidence: 99%

An Ensemble of Optimal Trees for Class Membership Probability Estimation

Khan

Gul

Mahmoud

et al. 2016

Analysis of Large and Complex Data

Self Cite

View full text Add to dashboard Cite

show abstract

“…Table 2 shows the 15 selected features from the entire features space extracted from the trace file that are generated in ns-2. POS is considered the most suitable and efficient scheme with a dataset that has common classification problems such as outliers and high-dimensional binary [36,37]. It was employed to calculate the overlapping rate in the features extracted.…”

Section: Feature Sets and Extractionmentioning

confidence: 99%

“…It was employed to calculate the overlapping rate in the features extracted. The R code the POS method using the following pseudocode [36]:…”

Section: Feature Sets and Extractionmentioning

confidence: 99%

Intelligent Intrusion Detection of Grey Hole and Rushing Attacks in Self-Driving Vehicular Networks

2016

View full text Add to dashboard Cite

Vehicular ad hoc networks (VANETs) play a vital role in the success of self-driving and semi self-driving vehicles, where they improve safety and comfort. Such vehicles depend heavily on external communication with the surrounding environment via data control and Cooperative Awareness Messages (CAMs) exchanges. VANETs are potentially exposed to a number of attacks, such as grey hole, black hole, wormhole and rushing attacks. This work presents an intelligent Intrusion Detection System (IDS) that relies on anomaly detection to protect the external communication system from grey hole and rushing attacks. These attacks aim to disrupt the transmission between vehicles and roadside units. The IDS uses features obtained from a trace file generated in a network simulator and consists of a feed-forward neural network and a support vector machine. Additionally, the paper studies the use of a novel systematic response, employed to protect the vehicle when it encounters malicious behaviour. Our simulations of the proposed detection system show that the proposed schemes possess outstanding detection rates with a reduction in false alarms. This safe mode response system has been evaluated using four performance metrics, namely, received packets, packet delivery ratio, dropped packets and the average end to end delay, under both normal and abnormal conditions.

show abstract

“…These features dramatically decrease the prediction performance of the algorithms (Nettleton et al (2010)). Feature selection methods that investigate the most discriminative features from the original features are usually recommended to mitigate the effect of such non-informative features (Mahmoud et al 2014a(Mahmoud et al , 2014b. However, different feature selection methods result in different feature subsets for the same data set thus varying feature relevancy.…”

Section: Introductionmentioning

confidence: 99%