Nowadays, computer networks and the Internet are unprotected from many security threats. Introducing adaptive and flexible security-related techniques is challenging because of the new types of frequently occurring attacks. An intrusion detection system (IDS) is a security device similar to other measures, including firewalls, antivirus software, and access control models devised to strengthen communication and information security. Network intrusion detection system (NIDS) plays a vital function in defending computer networks and systems. However, several issues concerning the sustainability and feasibility of existing techniques are faced with recent networks. These concerns are directly related to the rising levels of necessary human interactions and reducing the level of detection accuracy. Several approaches are designed to detect and manage various security threats in a network. This study uses Chimp Chicken Swarm Optimization-based Deep Long Short-Term Memory (ChCSO-driven Deep LSTM) for the intrusion detection process. A CNN feature extraction process is necessary for effective intrusion detection. Here, the Deep LSTM is applied for detecting network intrusion, and the Deep LSTM is trained using a designed optimization technique to enhance the detection performance.

Due to the increase in the use of the internet all over the world for business and education activates, cybercrime is increasing day by day in spite of the development of security protocols and algorithms. Recent research is based on the intrusion detection system. We attempt to develop is a secure protocol to detect malicious data, along with actual data, in the incoming data traffic. An intrusion detection system based on a recurrent neural network (RNN) classifier for feature reduction. Failure in intrusion detection apparatus results in a series of negatives ranging from loss of confidential data and thereof reducing reliability for the end-user. Hence, detection systems play an important role in the service end user. In the proposed work, an intelligent system is developed based on machine learning techniques specifically RNN wherein, a novel algorithm is developed for combining a correlation and information gain, feature reduction is achieved. A feed-forward neural network is then fed these reduced features for testing and training on the NSL-KDD dataset. Normally, pre-processing of the dataset is carried out before the training phase. It helps us to regularize instances of each class in the dataset. Attack and non-attack classes are formed which helps us to implement this developed algorithm for giving the best results in terms of Feature ranking. Our algorithms reduced the number of features, which in turn reduced in preprocessing time of extracting features related to information gain and correlation in the dataset.

No abstract

The following paper provides a novel approach for Network Intrusion Detection System using Machine Learning and Deep Learning. This approach uses two MLP (Multi-Layer Perceptron) models one having 3 layers and other having 6 layers. Random Forest is also used for classification. These models are ensembled in such a way that the final accuracy is boosted and also the testing time is reduced. Researchers have implemented various ways for the ensemble of multiple models but we are using contradiction management concept to ensemble machine learning models. Contradiction Management concept means if two machine learning models are contradicting in their decisions (in our case 3-layer MLP and Random Forest), then the third model’s (6-layer MLP) decision is considered whose accuracy is higher than the previous models. The third model is only used for testing when the previous two models contradict in their decision because the testing time of third model is higher than the two previous models as the third model has complex architecture. This approach increased the final accuracy as ensemble of multiple models is done and also testing time has reduced. The novelty of this paper is the choice and the combination of the models for the purpose of Network security.

In the current times, cyber-attacks are becoming more sophisticated and modern; this has increased the threat in precisely detecting intrusion. Inability to restrict the intrusions can degrade the validity of security administrations and loss of data confidentiality, integrity, and availability. Hence, Detection is an important step in avoiding such attacks; once an issue is detected properly, effective countermeasures can be deployed. Intrusion Detection Systems (IDS) plays a very crucial role and help to detect incoming attacks. Network-based IDS is an important tool used to protect the computer network against malicious attacks and threats. An application of the Bayesian Information Gain concept for feature selection and Deep Recurrent neural network (Deep RNN) to model building is proposed in this paper to increase the efficiency of a network intrusion detection system. The Bayesian Information Gain concept is used to select important features, which have high predictive power. Deep RNN classifier successfully plays out the intrusion detection system measure utilizing the hidden layers dependent on the weight and bias-related with the classifier. Appropriately, the Adam optimization algorithm to build the precision of the model ideally tunes the weights and bias.