An Intelligent Approach for Intrusion Detection using Convolutional Neural Network

Intrusion Detection Systems (IDS) is one of the important aspects of cyber security that can detect the anomalies in the network traffic. IDS are a part of Second defense line of a system that can be deployed along with other security measures such as access control, authentication mechanisms and encryption techniques to secure the systems against cyber-attacks. However, IDS suffers from the problem of handling large volume of data and in detecting zero-day attacks (new types of attacks) in a real-time traffic environment. To overcome this problem, an intelligent Deep Learning approach for Intrusion Detection is proposed based on Convolutional Neural Network (CNN-IDS). Initially, the model is trained and tested under a new real-time traffic dataset, CSE-CIC-IDS 2018 dataset. Then, the performance of CNN-IDS model is studied based on three important performance metrics namely, accuracy / training time, detection rate and false alarm rate. Finally, the experimental results are compared with those of various Deep Discriminative models including Recurrent Neural network (RNN), Deep Neural Network (DNN) etc., proposed for IDS under the same dataset. The Comparative results show that the proposed CNN-IDS model is very much suitable for modelling a classification model both in terms of binary and multi-class classification with higher detection rate, accuracy, and lower false alarm rate. The CNN-IDS model improves the accuracy of intrusion detection and provides a new research method for intrusion detection.

A Conundrum of Peer Evaluation of Predicting Poor Prognostic Models Affecting the Mortality Rate of Covid -19

Mankind is confronting these days a histrionic pandemic scene with the Coronavirus proliferation over all continents. The Covid-19 pandemic outbreak is as yet not very much portrayed, and numerous research teams everywhere on the world are chipping away at one or the other restorative therapeutic issues or immunization issues. The outburst of COVID-19 has constituted a danger to wellbeing of world. With the expanding number of individuals tainted, medical services frameworks, particularly those in economically emerging nations, are bearing gigantic pressing factor for the devising a prognostic model. There is a dire requirement for the analysis of COVID-19 and the anticipation of inpatients. To diminish these issues, a data statistical information driven clinical aid framework is advanced in this paper. In view of two real world datasets in Wuhan, China, the proposed framework coordinates information from various sources with tools of Machine Learning (ML) to anticipate COVID-19 tainted likelihood of suspected patients in their first visit, and afterward foresee mortality of affirmed cases. As opposed to picking an interpretable calculation, this framework isolates the clarifications from ML models. It can do help to patient triaging and give some valuable guidance to specialists and doctors. A prognosis model is in the way of extraordinary premium for specialists to adjust their consideration methodology for therapeutic or diagnosis procedure.

An Estimation and Evaluation of Network Availability in Link State Routing Networks

The internet is taking component in a developing feature in every non-public and professional activity. The real-time, delay sensitive and mission-essential purposes, community availability requirement is beforehand for internet carrier providers (ICPs). The loop-loose criterion (LLC) approach has been extensively deployed through numerous ICPs for handling the best network component failure state of affairs in fantastic internet through. The achievement of LLC lies in its inherent simplicity; however, this comes at the rate of letting certain failure. To reap complete failure safety with LLC without incurring significant extra, a singular link protection scheme, hybrid hyperlink protocol (HLP), to reap failure routing. In contrast with in advance schemes, HLP guarantees tall network in a greater surroundings pleasant way. HLP is carried out in stages. Initial level substances a surroundings pleasant LLC primarily based totally on (MNP-e). The complexity of the set of rules is decrease than that of Dijkstra’s set of rules and might gift similar to network availability with LCC (Loop-loose criterion). Moment level substances backup direction safety based on MNP-e, the area totally a minimum type of need to be protected, to fulfill the network requirement. We don't forget those algorithms in a massive spread of associated, real and actual, and the effects display that HLP can achieve lofty network without introducing apparent.

Intrusion Detection Using Multilayer Perceptron in Software Defined Network

The advancement of cloud and IoT technologies, has made network administration more difficult. Software-Defined Networking is one of the trending technologies which replaces the traditional networking domain with the programmable network configuration. In the current development of the network architecture, data security plays a prominent role. Many strategies for dealing with network attacks have been developed, among them deep learning is one of the most advanced technology. The paper aims to classify the network traffic into normal traffic and attack traffic with Multilayer Perceptron (MLP). The simulation uses a python programming language with many packages like Numpy, sci-kit, seaborn, etc. in a mininet SDN test bed with the Ryu controller. From the obtained results proposed algorithm gives better accuracy for classifying the attack traffic and normal traffic in the network.

A Novel Framework for Sign Language to Text Conversion using Convolutional Neural Network

Sign language is the only way of method to communication for hearing impaired and deaf-dumb peoples. The system will recognize the signs between signers and non-signers, this will give the meaning of sign. The proposed method is helpful for the people who have hearing difficulties and in general who use very simple and effective method is sign language. This system can be used for converting sign language to text using CNN approach. An image capture system is used for sign language conversion. It captures the signs and display on the screen as writing. Results prove that the planned methodology for sign detection is more effective and has high accuracy. Experimental results will acknowledge the signs that the planned system is 80% accuracy.

Review of Ensemble-Based Filter Feature Selection Techniques for Building Intrusion Detection System

Intrusion detection systems (IDS) play a critical role in network security by monitoring network traffic for malicious activities and detecting vulnerability exploits against target applications or computers. A large number of redundant and irrelevant features increase the dimensionality of the dataset, which increases the computational overhead on the system and reduces its performance. This paper studies different filter-based feature selection techniques to improve performance of system. Feature selection techniques are used to select a well performing subset of features followed by technique of ensemble learning, which selects an optimal subset of features by combining multiple subsets of features. Feature selection combined with ensemble learning is explored in this paper. The performance of the algorithms implemented in existing research in terms of accuracy, false alarm rates, and true positive rates is explored, and their shortcomings are observed.

Lung Cancer Classification Techniques

Diagnosis of lung cancer with high accuracy rate is most difficult task to make remarkable vary in survival rate of patients. Different imaging techniques are used by radiologists and specialists to diagnose lung cancer such as Computer tomography (CT), X-ray and Magnetic Resonance Imaging (MRI). These methods help us to predict the malignant or benign or normal nodules present in the lungs. This proposed work is to build a lung classification system that can classify the images as malignant or benign or normal cases and give best accuracy for predicting lung cancer. In this “IQ_OTH/NCCD” lung cancer dataset is used which consist of total 1190 images of lung CT scans slices for 110 cases. CT scans in DICOM formats is utilized in this research work. In this proposed work by applying machine learning techniques such as Artificial Neural Network (ANN) and Convolutional Neural Network (CNN), classify the malignant or normal or benign lung nodule cases and finally compare all the attained results. This work finds the accuracy of applied classification systems and finally CNN model outperforms with an accuracy of 98%. Accuracy of ANN model is observed to be 71%.