No abstract

The banking industry has been seeking novel ways to leverage database marketing efficiency. However, the nature of bank marketing data hindered the researchers in the process of finding a reliable analytical scheme. Various studies have attempted to improve the performance of Artificial Neural Networks in predicting clients’ intentions but did not resolve the issue of imbalanced data. This research aims at improving the performance of predicting the willingness of bank clients to apply for a term deposit in highly imbalanced datasets. It proposes enhanced Artificial Neural Network models (i.e., cost-sensitive) to mitigate the dramatic effects of highly imbalanced data, without distorting the original data samples. The generated models are evaluated, validated, and consequently compared to different machine-learning models. A real-world telemarketing dataset from a Portuguese bank is used in all the experiments. The best prediction model achieved 79% of geometric mean, and misclassification errors were minimized to 0.192, 0.229 of Type I & Type II Errors, respectively. In summary, an interesting Meta-Cost method improved the performance of the prediction model without imposing significant processing overhead or altering original data samples.

JavaScript obfuscation is a deliberate act of making a script difficult to understand by concealing its purpose. The prevalent use of obfuscation techniques to hide malicious codes and to preserve copyrights of benign scripts resulted in (i) missing detection of malicious scripts that are obfuscated and (ii) raising false alarms due to the benign scripts that are obfuscated. Automatic detection of obfuscated JavaScript is generally undertaken by tackling the problem from the readability perspective. Recently, Microsoft research team analyzed different levels of context‐based features to distinguish obfuscated malicious scripts from obfuscated benign ones. In this work, we raise the issue of existing readable versions of obfuscated scripts. Further, we discuss the challenges posed by readably obfuscated scripts against both JavaScript malware detectors and obfuscated scripts detectors. Therefore, we propose JavaScript Obfuscation Detector (JSOD), a completely static solution to detect obfuscated scripts including readable patterns. To evaluate JSOD, we compare it to the state‐of‐the‐art approaches to detect obfuscated malicious and obfuscated benign script, namely, Zozzle and Nofus. Our experimental results demonstrate the importance to detect readably obfuscated scripts and their sophisticated variations. Furthermore, they also show the superiority of JSOD approach against all relevant solutions. Copyright © 2014 John Wiley & Sons, Ltd.

Recently, spam on online social networks has attracted attention in the research and business world. Twitter has become the preferred medium to spread spam content. Many research efforts attempted to encounter social networks spam. Twitter brought extra challenges represented by the feature space size, and imbalanced data distributions. Usually, the related research works focus on part of these main challenges or produce black-box models. In this paper, we propose a modified genetic algorithm for simultaneous dimensionality reduction and hyper parameter optimization over imbalanced datasets. The algorithm initialized an eXtreme Gradient Boosting classifier and reduced the features space of tweets dataset; to generate a spam prediction model. The model is validated using a 50 times repeated 10-fold stratified cross-validation, and analyzed using nonparametric statistical tests. The resulted prediction model attains on average 82.32% and 92.67% in terms of geometric mean and accuracy respectively, utilizing less than 10% of the total feature space. The empirical results show that the modified genetic algorithm outperforms Chi 2 and P CA feature selection methods. In addition, eXtreme Gradient Boosting outperforms many machine learning algorithms, including BERT-based deep learning model, in spam prediction. Furthermore, the proposed approach is applied to SMS spam modeling and compared to related works.

No abstract

Deliverable and course project become the preferred mean to measure learner competency and attainment of intended learning outcomes in IT-fields. Proper setup and evaluation for teamwork projects remains a key challenge for e-learning systems. This study investigates the possibility to improve the early prediction of academic software engineering project failure by treating teamwork differently according to the distribution of teamwork participants. Two configurations of teamwork distribution are considered. In the first configuration, a teamwork may include international participants, but all team participants are affiliated to the same institution, namely local teamwork. In the second configuration, a teamwork may include participants from different institutions, namely global teamwork. Software engineering projects are approached from two distinct perspectives. First, obeying the best practices during the system development life cycle (SDLC), namely, process perspective. Second, characteristics of the final deliverable deployed at each milestone of the SDLC, namely product perspective. A publicly released dataset collected by a designated e-learning environment is leveraged to validate the proposed approach. Results indicate a noticeable variance among local and global distributions. These results puts evidence that the reasons behind software engineering teamwork project failures may vary depending on the distribution of the teamwork, local vs. global. Consequently, it advise to customize e-learning systems differently according to the teamwork distribution.

It is critical to safeguard confidential data, especially secret and private messages. This study introduces a novel data cryptography approach. The new approach will be capable of encrypting and decrypting any communication size. The suggested approach will use a sophisticated private key with a convoluted structure. The private key will have 5 components with a double data type to prevent guessing or hacking. The confidential data will produce two secret keys, the first of which will be taken from the image key. These keys will be vulnerable to slight changes in private key information. To maximize the approach's efficiency, the suggested method will deal with lengthy messages by splitting them into chunks. On the other hand, the chaotic logistic map model will be used to create the second key. The suggested technique will be implemented, and several sorts of analysis (sensitivity, quality, security, and speed analysis) will be undertaken to demonstrate the benefits of the proposed method. The quality metrics MSE, PSNR, and CC will be computed to validate the suggested method's quality. To illustrate the efficiency of the proposed technique, encryption and decryption times will be measured, and cryptography throughputs will be determined. Various PKs will be tried throughout the decryption process to demonstrate how sensitive the produced outputs are to changes in the private key. The suggested approach will be tested, and the results will be compared to the results of existing methods to demonstrate the improvement offered by the proposed method.

Colored digital images are one of the most important types of digital data to be used in many vital applications, which require a safe way to protect them from hacking operations and the danger of intruders and data thieves. This paper presents an effective and safe method for storing digital colored images (CASDC). A high level of protection is provided through a complex secret key agreed upon between the sender and the receiver. The secret key consists of nine decimal digits (and can be increased as needed). These digits are processed to extract three values for each color of the three color channels. A left rotation process is performed for the value of each color to produce three new values, where an exclusion process is performed between them to obtain the encrypted value for the color. CASDC is evaluated against a wide range of images to calculate its throughput to show the extent to which this method fulfills encryption and decryption requirements. The Mean Square Error (MSE) values, Peak Signal to Noise Ratio (PSNR), and Correlation Coefficient for the three primary channels of the RGB coloring system were analyzed. The practical results of the proposed method are compared with other standard methods such as Data Encryption Standard (DES), Tripple-DES (3DES), Advanced Encryption Standard (AES), and Blow Fish (BF). According to the obtained results, CASDC outperforms all standard methods in terms of efficiency by reducing the time of encryption and decryption and increasing the throughput of the corresponding process. Besides, CASDC is robust against breaks, as the attempts to break the private key will require hundreds of years in the best case.