Search

Due to the rapid development of network technology in recent years, network security has drawn a lot of attention. Intrusion detection systems are crucial in preventing unwanted traffic from entering networks and computer systems. For network security, it is essential to increase the detection accuracy of network attacks using a range of strategies. However, to increase the detection accuracy of DDoS attacks, the existing model uses a Support vector machine (SVM) and K-Nearest Neighbors (KNN), which does not address the misclassification of data during transmission. Hence, a novel DDoS attack Over Flash Crowd Using Cross GAN (XGAN) has been proposed to classify the performance by enhancing the detection of DDoS attacks in the network which utilizes the information gain, chi-square, and gain ratio to determine the features first using a wrapper-based feature selection ensemble. There is no data collection available right now that has both flash crowd and DDoS sample data. To achieve more accurate categorizations using any classification model, a Generative Adversarial Network (GAN) technique is used to mimic both in the same data set. Then, a Cross Generative Adversarial Network (XGAN), a mixture of two sets of GANs that construct and classify even the imitation damaging attacks with high accuracy, has been provided to improve the detection performance of the model by minimizing the imbalance of attack records. Hence the proposed methodology enhanced the DDoS detection with high accuracy. [ABSTRACT FROM AUTHOR]

DDoS (distributed denial of service) attack is a constant and serious danger to the Internet. New application-layer-based DDoS attacks that use valid requests to overload target resources are more undetected than low-layer DDoS attacks. When flash crowd assaults imitate or occur during a flash crowd event on webservers, the situation may become more dangerous. Flash Crowd Assaults (FCAs) are DDoS attacks that flood victim services, such as Web servers, with well-formed requests created by a large number of bots. Because both valid and attack requests appear identical, it's difficult to identify and filter such attacks.Hence, differentiating DDoS and FC is a critical task in protecting web servers against attacks, which can be fatal to cyber-systems.In this article, a novel Deep Neural Network empowered Bi-Directional Cross Generative Adversarial Network (GAN) is introduced for recognizing and separating distributed denial-of-service attacks from flash crowd attacks over web applications. The suggested research uses an ensemble feature selection approach to construct real samples of attack data in order to accomplish the goal. In the meanwhile, the generator network uses a bidirectionally cross GAN to generate bogus attack data by observing the random noise vector as input. The discriminator receives both manual and model input and using DNN to distinguish between DDoS assaults and flash crowd attacks. The proposed model is implemented in the working platform of python and is tested using the evaluation metrics like accuracy, precision, recall, and f1-measure.From the implementation result, is evident that the proposed model achieves an effective performance in whichthe obtained accuracy of the proposed approach is 96.58% which is comparatively higher than the existing techniques. [ABSTRACT FROM AUTHOR]

With the increasing number of telecom providers and services, churn prediction gains tremendous interest in the current decade. The prediction models based on machine learning are the greatest fortune for customer retention campaigns as they pave the way to predict potential churners. Though the prediction is accurate by applying the machine learning classifiers, there are huge concerns imputing complexity in the prediction process. In most cases, the churn prediction becomes complex due to the telecom data's inconsistency, sparsity, and hugeness. An effective churn prediction model is proposed to tackle such kinds of issues,. The proposed churn prediction is based on the optimized deep learning classifier built in the spark architecture to handle the hugeness of the telecom data. The optimized deep learning classifier is established through the optimal training of the deep convolutional neural network (DCNN) using the proposed Firefly-Spider Optimization (FSO), which is the integration of Spider Monkey Optimization (SMO) and firefly optimization algorithm (FA). The proposed prediction model's effectiveness is analyzed using the Churn in Telecom's dataset based on the performance measures. The proposed prediction model acquired the maximal dice coefficient, accuracy, and Jaccard coefficient of 94.61%, 94.76%, and 94.80%. [ABSTRACT FROM AUTHOR]

The churn prediction based on telecom data has been paid great attention because of the increasing the number telecom providers, but due to inconsistent data, sparsity, and hugeness, the churn prediction becomes complicated and challenging. Hence, an effective and optimal prediction of churns mechanism, named adaptive firefly-spider optimization (adaptive FSO) algorithm, is proposed in this research to predict the churns using the telecom data. The proposed churn prediction method uses telecom data, which is the trending domain of research in predicting the churns; hence, the classification accuracy is increased. However, the proposed adaptive FSO algorithm is designed by integrating the spider monkey optimization (SMO), firefly optimization algorithm (FA), and the adaptive concept. The input data is initially given to the master node of the spark framework. The feature selection is carried out using Kendall's correlation to select the appropriate features for further processing. Then, the selected unique features are given to the master node to perform churn prediction. Here, the churn prediction is made using a deep convolutional neural network (DCNN), which is trained by the proposed adaptive FSO algorithm. Moreover, the developed model obtained better performance using the metrics, like dice coefficient, accuracy, and Jaccard coefficient by varying the training data percentage and selected features. Thus, the proposed adaptive FSO-based DCNN showed improved results with a dice coefficient of 99.76%, accuracy of 98.65%, Jaccard coefficient of 99.52%. [ABSTRACT FROM AUTHOR]

Nowadays, organizations collect volumes of data for future analysis. Motivated by this amount of data and requirements of Web 2.0, a plethora of non-relational databases (NoSQL) emerged in recent years. However, several security features in relational databases (e.g., access control) have been left in non-relational management systems to be developed by the application, which can raise security breaches. This paper proposes two distinctive security model, based on the use of metadata, to provide access control for NoSQL graph-oriented database management system and organization based system. The goal is to support the development of applications that use graph-oriented database in preserving the integrity of stored data and protect them from non-authorized access. A case study has been performed as proof of concept, where the model has been instantiated and implemented for Neo4j database. A network management data security technique for executives based on NoSQL is suggested for better assurance of Organization based security systems. These are the two distinctive security models presented in this study. [ABSTRACT FROM AUTHOR]

Remote sensing techniques have been extensively utilized for recognition of land use and land cover structures. Land evidence can be definitely composed by classification of satellite images in the perspective of their practice. In this paper study area has been classified into five classes i.e. vegetation, agriculture, water body, open area and urban land by classification of fused images obtained from various fusion techniques. The spatial and spectral determinations of various satellite images make availableimprovedevidence with the encouragementof imageprocessing and image fusion of both multispectral and spatial images. The input images fused together are multispectral image and panchromatic images obtained from IRS-1D satellite utilizing LISS III. Matlab 10.0 software has been used for image processing, fusion and classification of the images. The Principal Component Analysis (PCA), wavelet transform, fuzzy and neuro fuzzy techniques arehave been used for image fusion. The resultant images have been classified using the supervised and unsupervised classification techniques;decision tree classifier and K-Meansalgorithms and evaluationconcerning them in standings of their accuracy. [ABSTRACT FROM AUTHOR]

Frequent itemsets plays a crucial role in many data mining tasks that aimed to find interesting patterns. Frequent Pattern mining is used for find frequent itemsets among items in a given data set and the exact frequency of each itemset. An objective of frequent pattern mining is to develop a systematic method for the given data set and find frequent items. Obtained frequent items can be used for enhancing the business return of interest. The Synoptic/Concise representation of frequent Itemsets sacrifice/presents readability and direct interpretability by a data analyst of the concise patterns extracted/picked up. This paper presents, FRI Miner, a tool for Frequent Regular Itemsets Mining, designed to evaluate RegularMine and FPGrowth, and experimented on real time datasets. [ABSTRACT FROM AUTHOR]

Spatial databases store information related to objects positional locations e.g., various kinds of multidimensional objects represented by points, line segments, polygons and other kinds of geometric entities. These databases should support for efficient storage, indexing and querying of spatial data. Special purpose indexing structures are important for accessing the spatial data, and for processing spatial queries. To handle the spatial data efficiently, a database system needs an indexing mechanism that will help it to retrieve data items quickly according to their spatial locations instead of using GIS (Geospatial Information System). Hence, in this paper, authors evaluated R, R+ and R* dynamic indexing algorithms, on a variety of real time and synthetic datasets, to access the spatial data. From the experimental study, one can observe that the R* indexing technique performs well (than R and R+) for spatial databases. [ABSTRACT FROM AUTHOR]