Zeynep Orman

A novel feature-selection approach based on the cuttlefish optimization algorithm for intrusion detection systems

A modified version of the cuttlefish algorithm is discussed.The proposed model can be used as a novel feature-selection model.Cuttlefish algorithm is used as a search strategy to find optimal subset of features.Decision tree is used to evaluate the quality of the selected features.Data pre-processing for feature selection is also examined in the paper. This paper presents a new feature-selection approach based on the cuttlefish optimization algorithm which is used for intrusion detection systems (IDSs). Because IDSs deal with a large amount of data, one of the crucial tasks of IDSs is to keep the best quality of features that represent the whole data and remove the redundant and irrelevant features. The proposed model uses the cuttlefish algorithm (CFA) as a search strategy to ascertain the optimal subset of features and the decision tree (DT) classifier as a judgement on the selected features that are produced by the CFA. The KDD Cup 99 dataset is used to evaluate the proposed model. The results show that the feature subset obtained by using CFA gives a higher detection rate and accuracy rate with a lower false alarm rate, when compared with the obtained results using all features.

New sufficient conditions for global stability of neutral-type neural networks with time delays

This paper studies the equilibrium and stability properties of the class of neutral-type neural network model with discrete time delays. By employing a Lyapunov functional and examining the time derivative of the Lyapunov functional, we obtain some delay independent sufficient conditions for the existence, uniqueness and global asymptotic stability of the equilibrium point for this class of neutral-type systems. The obtained conditions can be easily verified as they can be expressed in terms of the network parameters only. We also compare our results with the previous corresponding results derived in the literature by giving some numerical examples.

New results for global stability of Cohen-Grossberg neural networks with multiple time delays

This paper studies the global convergence properties of Cohen-Grossberg neural networks with multiple time delays. Without assuming the symmetry of interconnection weight coefficients, and the differentiability and boundedness of activation functions, and by employing Lyapunov functionals, we derive new delay independent sufficient conditions under which a delayed Cohen-Grossberg neural network converges to a unique and globally asymptotically stable equilibrium point. Several examples are given to illustrate the advantages of our results over the previously reported results in the literature.

An Analysis of Stability of a Class of Neutral-Type Neural Networks with Discrete Time Delays

The problem of existence, uniqueness, and global asymptotic stability is considered for the class of neutral-type neural network model with discrete time delays. By employing a suitable Lyapunov functional and using the homeomorphism mapping theorem, we derive some new delay-independent sufficient conditions for the existence, uniqueness, and global asymptotic stability of the equilibrium point for this class of neutral-type systems. The obtained conditions basically establish some norm and matrix inequalities involving the network parameters of the neural system. The main advantage of the proposed results is that they can be expressed in terms of network parameters only. Some comparative examples are also given to compare our results with the previous corresponding results and demonstrate the effectiveness of the results presented.

New results for global stability of cohen-grossberg neural networks with discrete time delays

This paper studies the global convergence properties of Cohen-Grossberg neural networks with discrete time delays. Without assuming the symmetry of interconnection weight coefficients, and the monotonicity and differentiability of activation functions, and by employing Lyapunov functionals, we derive new delay independent sufficient conditions under which a delayed Cohen-Grossberg neural network converges to a globally asymptotically stable equilibrium point. Some examples are given to illustrate the advantages of the results over the previously reported results in the literature.

An improved stability result for delayed Takagi–Sugeno fuzzy Cohen–Grossberg neural networks

This work proposes a novel and improved delay independent global asymptotic stability criterion for delayed Takagi-Sugeno (T-S) fuzzy Cohen-Grossberg neural networks exploiting a suitable fuzzy-type Lyapunov functional in the presence of the nondecreasing activation functions having bounded slopes. The proposed stability criterion can be easily validated as it is completely expressed in terms of the system matrices of the fuzzy neural network model considered. It will be shown that the stability criterion obtained in this work for this type of fuzzy neural networks improves and generalizes some of the previously published stability results. A constructive numerical example is also given to support the proposed theoretical results.

A DIDS Based on The Combination of Cuttlefish Algorithm and Decision Tree

Different Distributed Intrusion Detection Systems (DIDS) based on mobile agents have been proposed in recent years to protect computer systems from intruders. Since intrusion detection systems deal with a large amount of data, keeping the best quality of features is an important task in these systems. In this paper, a novel DIDS based on the combination of Cuttlefish Optimization Algorithm (CFA) and Decision Tree (DT) is proposed. The proposed system uses an agent called Rule and Feature Generator Agent (RFGA) to generate a subset of features with corresponding rules. RFGA agent uses CFA to search for optimal subset of features, while DT is used as a measurement on the selected features. The proposed model is tested on the KDD Cup 99 dataset. The obtained results show that the proposed system gives a better performance even with a small subset of 5 features when compared with using all 41 features.