Convolutional neural network (CNN) is a basic configuration of neural networks that can perform deep learning. There are many applications based on CNN in fields of image processing, machine learning, and data analysis. The CNN is a complex neural network with a various number of hidden layers (depth) and a large number of neurons. The depth of the CNN is the essential factor that determines how the network can perform a complicated task. The design phase of CNNs requires potential from non-experts of machine learning. In this chapter, a fully automated algorithm to develop CNN was proposed based on firefly optimization. The proposed method can design a CNN structure with any number of layer depth without any limitation on the depth value. The proposed method employed the skip connection as a fundamental building block of CNN. A modified firefly algorithm was presented base on the $$k-$$nearest neighbor attraction model to reduce the computational complexity of the firefly. The CIFAR-10 and CIFAR-100 were used for the training and validation of the proposed method to perform image classification. The proposed method provided high accuracy when compared to the cutting-edge approaches.

Automatic human iris verification is an active research area with numerous applications in security purposes. Unfortunately, most of feature extraction methods in human iris verification systems are sensitive to noise, scale and rotation. This paper proposes an integrated hybrid model among Discrete Wavelet Transform, Wavelet Neural Network and Genetic Algorithms for optimizing the feature extraction and verification methods. For any iris image, the wavelet features are extracted by Discrete Wavelet Transform without any dependency on scale and pixels' intensity. Besides, Wavelet Neural Network classifier is integrated as a local optimization method to solve the orientation problem and increase the intrinsic features. In solving the down sample process caused by DWT, each human iris should be characterized by a set of parameters of its optimal wavelet analysis function at a determined analysis level. Thus, distributed Genetic Algorithms, meta-heuristic algorithm, is introduced as a global optimization searching technique to discover the optimal parameter values. The details and limitation of this paper will be discussed where a comparative study should appear. Moreover, conclusions and future work are described.

Handling uncertainty and vagueness in real world becomes a necessity for developing intelligent and efficient systems. Based on the credibility theory, a fuzzy clustering approach that improves the classification accuracy is targeted by this work. This paper introduces a design of an efficient set of fuzzy rules that are inferred by a hybrid model of SOFM (Self Organized Features Maps) and FRANTIC-SRL (Fuzzy Rules from ANT-Inspired Computation - Simultaneous Rule Learning). Self-Organized Features Maps cluster inputs using self-adaption techniques. They are useful in generating fuzzy membership functions for the subsets of the fuzzy variables. The generated fuzzy variables are ranked by means of the credibility measure wherever the weighted average of their confidence level is determined. FRANT IC-SRL builds the fuzzy classification rule set using the ranked credibility variables in a simultaneous process. Moreover, the whole fuzzy system is evaluated based on the credibility value. The details and limitations of the proposed model are illustrated. Also, the experimental results and a comparison with previous techniques in generating fuzzy classification rules from medical data sets are declared.

Handwritten character recognition is one of the focused areas of research in the field of Pattern Recognition. In this paper, a hybrid model of rough neural network has been developed for recognizing isolated Arabic/Farsi digital characters. It solves the neural network problems; proneness to overfitting, and the empirical nature of model development using rough sets and the dissimilarity analysis. Moreover the perturbation in the input data is violated using rough neuron. This paper describes an evolutionary rough neural network based technique to recognize Arabic/Farsi isolated handwritten digital characters. This method involves hierarchical feature extraction, data clustering and classification. In contrast with conventional neural network, a comparative study is appeared. Also, the details and limitations are discussed.

This paper presents some of the results of our probabilistic cellular automaton (PCA) based epidemic model. It is shown that PCA performs better than deterministic ones. We consider two possible ways of interaction that relies on a two- way split rules either horizontal or vertical interaction with 2 different probabilities causing more of the best possible choices for the behavior of the disease. Our results are a generalization of that Hawkins et al done. Because of the spread of diseases, a technical innovative model should be made to recover their time regions. Many researches tried to solve this problem based on medical dis- ease feature, which suffer from unpredictable ones. Whilst a single infected host might not be significant, a disease that spreads through a large population yields serious health and economic threats. In this sense, mathematical epidemiology is concerned with modeling the spread of infectious disease in a population (see 2). The aim is generally to understand the time course of the disease with the goal of controlling its spread. Traditionally, the majority of existing mathematical models to simulate epidemics are based on ordinary differential equa- tions. These models have serious drawbacks in that they ne- glect the local characteristics of the spreading process and they do not include variable susceptibility of individuals. Spe- cifically, they fail to simulate in a proper way (1) the individ- ual contact processes, (2) the effects of individual behavior, (3) the spatial aspects of the epidemic spreading, and (4) the effects of mixing patterns of the individuals.

Because of the patient's inconsistent data, uncertain Thyroid Disease dataset is appeared in the learning process: irrelevant, redundant, missing, and huge features. In this paper, Rough sets theory is used in data discretization for continuous attribute values, data reduction and rule induction. Also, Rough sets try to cluster the Thyroid relation attributes in the presence of missing attribute values and build the Modified Similarity Relation that is dependent on the number of missing values with respect to the number of the whole defined attributes for each rule. The discernibility matrix has been constructed to compute the minimal sets of reducts, which is used to extract the minimal sets of decision rules that describe similarity relations among rules. Thus, the rule associated strength is measured.

The need to handle uncertainty and vagueness in real world becomes a necessity for developing good and efficient systems. Fuzzy rules and their usage in fuzzy systems help too much in solving these problems away from the complications of probability mathematical calculations. Fuzzy rules deals will words and labels instead of values of the variables. These labels are called variable’s subsets and needed to be prepared carefully to make sure that the fuzzy rules depend on accurate propositions. This research tries to design an efficient set of rules that is used later for inference by a hybrid model of Self Organized Features Maps and Parallel Genetic Algorithms. Self Organized Features Maps capabilities to cluster inputs using self adoption techniques have been very useful in generating fuzzy membership functions for the subsets of the fuzzy variables. Then the Parallel Genetic Algorithms use these membership functions along with the training data set to find the most fit fuzzy rule set from a number of initial sub populations according to the fitness function. The illustrations of the proposed model and its sub modules along with the experimental results and comparisons with previous techniques in generating rules from data sets are declared.

Many companies and organizations use firewalls to control the access to their network infrastructure. When processing packages, conflicts due to rule overlaps can occur within the filtering policy. To solve these conflicts most firewall implementation use a first matching strategy through the ordering of rules. This way each packet processed by the firewall is mapped to the decision of the rule with highest priority. This strategy introduces however new configuration errors such as shadowing of rules and redundancy lead to inaccurate results. In this paper new algorithm called range algorithm introduced to get the best case for solving conflict and shadowing problems. Also get result rules that is free inconsistency and finding rules that cause inconsistency.

Many companies and organizations use firewalls to control the access to their network infrastructure. When processing packages, conflicts due to rule overlaps can occur within the filtering policy. To solve these conflicts most firewall implementation use a first matching strategy through the ordering of rules. This way each packet processed by the firewall is mapped to the decision of the rule with highest priority. This strategy introduces however new configuration errors such as shadowing of rules and redundancy lead to inaccurate results. In this paper new algorithm called range algorithm introduced to get the best case for solving conflict and shadowing problems. Also get result rules that is free inconsistency and finding rules that cause inconsistency.

Recently machine learning-based Intrusion Detection systems (IDs) have been subjected to extensive researches because they can detect both misuse and anomaly. Most of existing IDs use all features in the network packet to look for known intrusive patterns. Some of these features are irrelevant or redundant. Rough Set Classification (RSC), a modern learning algorithm, is used to rank features extracted for detecting intrusions and generate intrusion detection models. In this paper a new hybrid model RSC-PGA (Rough Set Classification Parallel Genetic Algorithm) is presented to address the problem of identifying important features in building an intrusion detection system, increase the convergence speed and decrease the training time of RSC. Tests are done on KDD-99 dataset used for The Third International Knowledge Discovery and Data Mining Tools Competition. Results showed that the proposed model gives better and robust representation of rules as it was able to select features resulting in great data reduction, time reduction and error reduction in detecting new attacks. Keywords— Intrusion detection, Parallel genetic algorithm, Rough set classification.

The need for network security will become more urgent over time. Not only are cybercriminals constantly inventing new ways to steal electronic information, they are getting more nefarious in their use of it. An effective solution for many businesses is the firewall, which monitors and regulates traffic flow between public and private network environments to thwart attacks and prevent the entry of hostile code. Firewall ACLs can contain inconsistencies. There is an inconsistency if different actions can be taken on the same flow of traffic, depending on the ordering of the rules. Many approaches to solve inconsistency and redundancy problems last approach called range algorithm. In this paper analysis performance of the range algorithm will introduce due to range, time processing, CPU utilization, number of rules and accuracy.

ABSTRACT A new learning algorithm for space invariant Cellular Neural Network (CNN) is introduced. Learning is formulated as an optimization problem by combining rough sets and genetic programming. Rough Sets approach has been selected for creating priori knowledge about the actual effective cells, determining their significance in classifying the output, and discovering the optimal CNN structure. According to the lattice of CNN architecture and depending on the priori knowledge gained by rough sets, genetic programming will be used in deriving the cloning template. Exploration of any stable domain is possible by the current approach. Details of the algorithm are discussed and several application results are shown.

We purpose to find a new beneficial method for accelerating the Decision-Making and classifier support applied on imprecise data. This acceleration can be done by integration between Rough Sets theory, which gives us the minimal set of decision rules, and the Cellular Neural Networks. Our method depends on Genetic Algorithms for designing the cloning template for more accuracy. Some illustrative examples are given to demonstrate the effectiveness of the proposed method, whose advantages and limitations are also discussed.

A new learning algorithm for space invariant Uncoupled Cellular Neural Network is introduced. Learning is formulated as an optimization problem. Genetic Programming has been selected for creating new knowledge because they allow the system to find new rules both near to good ones and far from them, looking for unknown good control actions. According to the lattice Cellular Neural Network architecture, Genetic Programming will be used in deriving the Cloning Template. Exploration of any stable domain is possible by the current approach. Details of the algorithm are discussed and several application results are shown.

We propose a new technique based upon genetic programming to discover new learning rules for cellular neural networks. We choose genetic programming not only for its ability to discover the values of rule parameter but also for its ability to discover the optimal number of parameters and the form of the rules. A new supervised learning algorithm has been discovered and comparison with other different methods is taken into account.

This paper describes a hybrid framework for this kind of switching circuit. In this framework, Cellular Neural Networks (CNN) and rough sets are integrated into a hybrid system and used cooperatively during the system lifecycle. Rough sets and CNN were chosen for this application because they can discover patterns in ambiguous and imperfect data and provide tools for data and pattern analysis.