No abstract

Accurate identification of network intrusions is one of the biggest challenges of Network Intrusion Detection (NID) systems. In recent years Machine learning classification techniques have been used to precisely identify network intrusion. However, the multi class distribution in network intrusion detection system has found to be highly skewed, leading to classification accuracy problem due to class imbalance data set. The work presented in this paper not only explores the role of the attribute selection in improving classification accuracy but also investigates the problem of class imbalance using the Synthetic Minority Over-sampling (SMOTE) and under sampling of major classes. The classification performance is then evaluated over several types of classifiers. The outcome of this work is that for the class imbalance data set the under-sampling technique is more effective than SMOTE in detecting minor classes. It has also found during this research work that the decision tree algorithms (JRIP) and Naïve Bayes are more accurate classifiers as compared to the Radial basis neural network and support vector machine. However no single algorithm can be used for the classification of multiclass and it is proposed in this research work that combination of classifier consisting of Naïve Bayes and JRIP could be used for the classification of minor classes in an imbalance class data set of intrusion detection system.

Deep metric learning (DML) has achieved state-of-the-art results in several deep learning applications. However, this type of deep learning models has not been tested on the classification of electrical brain waves (EEG) for brain computer interface (BCI) applications. For the first time, we propose a triplet network to classify motor imagery (MI) EEG signals. Stockwell Transform has been used for converting the EEG signals in the time domain into the frequency domain, which resulted in improved DML classification accuracy in comparison to DML with Short Term Fourier Transform (0.647 vs. 0.431). DML model was trained with a topogram of concatenated 64 EEG channel spectrograms. The training batch was comprised of triplet pairs of the anchor, positive, and negative labeled epochs. The triplet network was able to train an embedding feature space that minimized the Euclidean distance between the embeddings of spectrograms of the same class and increased the distance between the embeddings of different labeled images. The proposed method has been tested on an EEG dataset of 109 untrained subjects. We showed that the DML classifier is able to converge with an extremely small number of training samples (~120 EEG trials) for only one subject per model, mitigating the well-known issue of the large inter-individual variability of human MI-BCI EEG which degrades the classification performance. The proposed preprocessing pipeline and the Triplet Network provide a promising method to classify MI-BCI EEG signals with much less training samples than the previous methods.

Abstract--The JPEG standard (IS O/ IEC 10918-1 ITU-T Recommendation T.81) defines compression techniques for image data. As a consequence, it allows to store and transfer image data with considerably reduced demand for storage space and bandwidth. From the four processes provided in the JPEG standard, only one, the baseline process is widely used. In this paper FPGA based High speed, low complexity and low memory implementation of JPEG decoder is presented. The pipeline implementation of the system, allow decompressing multiple image blocks simultaneously.The hardware decoder is designed to operate at 100MHz on Altera Cyclon II or Xilinx S partan 3E FPGA or equivalent. The decoder is capable of decoding Baseline JPEG color and gray images. Decoder is also capable of downscaling the image by 8. The decoder is designed to meet industrial needs. JFIF, DCF and EXIF standers are implemented in the design.

Radial basis function neural networks are used in a variety of applications such as pattern recognition, nonlinear identification, control and time series prediction. In this paper, the learning algorithm of radial basis function neural networks is analyzed in a feedback structure. The robustness of the learning algorithm is discussed in the presence of uncertainties that might be due to noisy perturbations at the input or to modeling mismatch. An intelligent adaptation rule is developed for the learning rate of RBFNN which gives faster convergence via an estimate of error energy while giving guarantee to the l 2 stability governed by the upper bounding via small gain theorem. Simulation results are presented to support our theoretical development.

Objectives It is not certain from current evidence which patient groups with non-visible haematuria (NVH) require urgent investigation and which investigations are sufficient. We report referral outcomes data from Scotland to identify patient groups who will benefit from urgent assessment to rule out urological cancer (UC) and whether full set of investigations are necessary in all referred patients. Materials and methods Data were collected from electronic patient records for patients referred with NVH to secondary care urology services between July 2017 and May 2020. The correlations between risk factors and final diagnosis were assessed using categorical variables in a multivariate logistic regression analysis and using chi-squared models. Statistical analysis was performed using IBM SPSS data editor version 25. Results Our study cohort comprised 525 patients (43.4% males; median age 66 years), in which UC was diagnosed in 25 patients (4.8%). Age > 60 years had sensitivity and NPV for UC of 92% and 99%, respectively. Univariate and multivariate analysis showed male sex, age ≥ 60 years and smoking were significant predictors of UC in patients with NVH (p < 0.05). There was no significant difference in UC in patients with history of LUTS, anticoagulation and previous UC. Conclusion The risk of urologic cancer in NVH patients is significant and male gender, age ≥ 60 years and smoking are significant predictors of UC. Patients with risk factors of UC require complete assessment of both the upper and lower urinary tract; however, in the absence of risk factors, patients do not require urgent or complete assessment.

No abstract

In the past, information safety was centered on event correlation designed for observing and spotting previously identified attacks. Due to the dynamic nature of multidimensional cyber-attacks, these models are no more acceptable. Specifically, these attacks use different strategies and procedures to find their way into and out of an organization. Traditional methods have reached their limit and thus new approaches are needed to find a solution for arising issues and challenges for big data security. To understand the current problem, we critically reviewed the literature related to big data security and the solutions proposed by the scientific community. In this paper, an ensemble approach for big data cybersecurity is proposed. To evaluate our approach, the given benchmark data is fed to three different classifiers namely to a k-nearest neighbor (KNN), support vector machine (SVM), multilayer perceptron (MLP) and the output of the single classifiers were compared to ensemble approach of the three classifiers. The reported results show that the ensemble approach for big data cybersecurity performs better than the single classifiers.