Mostafa A. Salama

Hybrid Intelligent Intrusion Detection Scheme

This paper introduces a hybrid scheme that combines the advantages of deep belief network and support vector machine. An application of intrusion detection imaging has been chosen and hybridization scheme have been applied to see their ability and accuracy to classify the intrusion into two outcomes: normal or attack, and the attacks fall into four classes; R2L, DoS, U2R, and Probing. First, we utilize deep belief network to reduct the dimensionality of the feature sets. This is followed by a support vector machine to classify the intrusion into five outcome; Normal, R2L, DoS, U2R, and Probing. To evaluate the performance of our approach, we present tests on NSL-KDD dataset and show that the overall accuracy offered by the employed approach is high.

Computational Social Networks: Security and Privacy

The continuous self-growing nature of social networks makes it hard to define a line of safety around these networks. Users in social networks are not interacting with the Web only but also with trusted groups that may also contain enemies. There are different kinds of attacks on these networks including causing damage to the computer systems and stealing information about users. These attacks are not only affecting individuals but also the organizations they are belonging to. Protection from these attacks should be performed by the users and security experts of the network. Advices should be provided to users of these social networks. Also security experts should be sure that the contents transmitted through the network do not contain malicious or harmful data. This chapter presents an overview of the social networks security and privacy issues and illustrates the various security risks and the tasks applied to minimize those risks. In addition, this chapter explains some of the common strategies that attackers often use and some possible counter measures against such issues.

Bi-Layer Behavioral-Based Feature Selection Approach for Network Intrusion Classification

Feature selection is a preprocessing step to machine learning, used to reduce the dimensionality of the dataset by removing irrelevant data. Variety of feature selection methods have been developed in the literature in order to increas the learning accuracy and reduce its complexity. In this paper we proposed a Bi-Layer behavioral-based feature selection approach. The proposed approach consists of two layers, in the first layer information gain is used to rank the features and select a new set of features depending on a global maxima classification accuracy. Then, in the second layer a new set of features is selected from within the first layer redacted data by searching for a group of local maximum classification accuracy in order to increase the number of reduced features. To evaluate the proposed approach it is applied on NSL-KDD dataset, where the number of features is reduced from 41 to 34 features in the first layer. Then reduced from 34 to 20 features in the second layer, which leads to improve the classification accuracy.

Deep Belief Network for clustering and classification of a continuous data

Deep Belief Network (DBN) is a deep architecture that consists of a stack of Restricted Boltzmann Machines (RBM). The deep architecture has the benefit that each layer learns more complex features than layers before it. DBN and RBM could be used as a feature extraction method also used as neural network with initially learned weights. The approach proposed depends on DBN in clustering and classification of continuous input data without using back propagation in the DBN architecture. DBN should have a better a performance than the traditional neural network due the initialization of the connecting weights rather than just using random weights in NN. Each layer in DBN (RBM) depends on Contrastive Divergence method for input reconstruction which increases the performance of the network.

Uni-class pattern-based classification model

This paper presents a model of a supervised machine learning approach for classification of a dataset. The model extracts a set of patterns common in a single class from the training dataset according to the rules of the pattern-based subspace clustering technique. These extracted patterns are used to classify the objects of that class in the testing dataset. The user-defined threshold dependence problem in this clustering technique has been resolved in the proposed model. Also this model solve the curse of dimensionality problem without the need of using a separate dimensionality reduction method. Another distinguishing point in this model is its dependence on the variation of the values of relative features among different objects. Experimental results on synthetic and real life datasets show that this approach is more efficient and effective than the existing techniques.

Employment of neural network and rough set in meta-learning

The selection of the optimal ensembles of classifiers in multiple-classifier selection technique is un-decidable in many cases and it is potentially subjected to a trial-and-error search. This paper introduces a quantitative meta-learning approach based on neural network and rough set theory in the selection of the best predictive model. This approach depends directly on the characteristic, meta-features of the input data sets. The employed meta-features are the degree of discreteness and the distribution of the features in the input data set, the fuzziness of these features related to the target class labels and finally the correlation and covariance between the different features. The experimental work that consider these criteria are applied on twenty nine data sets using different classification techniques including support vector machine, decision tables and Bayesian believe model. The measures of these criteria and the best result classification technique are used to build a meta data set. The role of the neural network is to perform a black-box prediction of the optimal, best fitting, classification technique. The role of the rough set theory is the generation of the decision rules that controls this prediction approach. Finally, formal concept analysis is applied for the visualization of the generated rules.

Rough Sets-Based Rules Generation Approach: A Hepatitis C Virus Data Sets

The risk of hepatitis-C virus is considered as a challenge in the field of medicine. Applying feature reduction technique and generating rules based on the selected features were considered as an important step in data mining. It is needed by medical experts to analyze the generated rules to find out if these rules are important in real life cases. This paper presents an application of a rough set analysis to discover the dependency between the attributes, and to generate a set of reducts consisting of a minimal number of attributes. The experimental results obtained, show that the overall accuracy offered by the rough sets is high.

Pattern-based subspace classification model

The use of patterns in predictive models has received a lot of attention in recent years. This paper presents a pattern-based classification model which extracts the patterns that have similarity among all objects in a specific class. This introduced model handles the problem of the dependence on a user-defined threshold that appears in the pattern-based subspace clustering. The experimental results obtained, show that the overall pattern-based classification accuracy is high compared with other machine learning techniques including Support vector machine, Bayesian Network, multi-layer perception and decision trees.

Reducing the influence of normalization on data classification

Principle Component Analysis (PCA) received a lot of attention over the past years and it is considered as a preprocessing method before many data mining models. PCA depends on the assumption that the input is normally distributed which is not true in many real life cases. On the other hand applying normalization on the input could change the structure of data and then affecting the outcome of multivariate analysis and calibration used in data mining. This paper introduces the effect of normalization methods before applying the conventional PCA. And It declares that the correlation PCA that uses the correlation matrix in PCA method could avoid such requirement. It proves that the correlation PCA leads to a better classification performance when the appropriate number of components is selected. The results also show that the resulted classification performance is independent on the normality of input.

Rough sets-based identification of heart valve diseases using heart sounds

Recently, heart sound signals have been used in the detection of the heart valve status and the identification of the heart valve disease. Heart sound data sets represents real life data that contains continuous and a large number of features that could be hardly classified by most of classification techniques. Feature reduction techniques should be applied prior applying data classifier to increase the classification accuracy results. This paper introduces the ability of rough set methodology to successfully classify heart sound diseases without the need applying feature selection. The capabilities of rough set in discrimination, feature reduction classification have proved their superior in classification of objects with very excellent accuracy results. The experimental results obtained, show that the overall classification accuracy offered by the employed rough set approach is high compared with other machine learning techniques including Support Vector Machine (SVM), Hidden Naive Bayesian network (HNB), Bayesian network (BN), Naive Bayesian tree (NBT), Decision tree (DT), Sequential minimal optimization (SMO).