Kevin Daimi’s research while affiliated with University of Detroit Mercy and other places

Many software systems contain large number of classes which make software testing very difficult. As software systems evolve, test suites become very large. It is usually very expensive to execute the entire test suites. Therefore, focusing the testing on the more error-prone classes helps in reducing the testing cost. In this work, we propose an approach for test case selection using software metrics. We examine the ability of several complexity and volume metrics to find the most complex and error-prone classes. Testers can then run test cases that are associated with the complex classes only. We focus our experiments on systems written in Java and tested with the JUnit testing framework. The results reveal that the proposed approach significantly reduce the number of test cases needed while detecting most of seeded errors.

Depression is a disorder characterized by misery and gloominess felt over a period of time. Some symptoms of depression overlap with somatic illnesses implying considerable difficulty in diagnosing it. This paper contributes to its diagnosis through the application of data mining, namely classification, to predict patients who will most likely develop depression or are currently suffering from depression. Synthetic data is used for this study. To acquire the results, the popular suite of machine learning software, WEKA, is used.

Depression is a disorder characterized by misery and gloominess felt over a period of time. Some symptoms of depression overlap with somatic illnesses implying considerable difficulty in diagnosing it. This paper contributes to its diagnosis through the application of data mining, namely classification, to predict patients who will most likely develop depression or are currently suffering from depression. Synthetic data is used for this study. To acquire the results, the popular suite of machine learning software, WEKA, is used.

Depression is a disorder characterized by misery and gloominess felt over a period of time. Some symptoms of depression overlap with somatic illnesses implying considerable difficulty in diagnosing it. This paper contributes to its diagnosis through the application of data mining, namely classification, to predict patients who will most likely develop depression or are currently suffering from depression. Synthetic data is used for this study. To acquire the results, the popular suite of machine learning software, WEKA, is used.

Cognitive radio network (CRN) is an evolving concept aiming at more efficiently exploiting the available spectrum for opportunistic network usage. Deploying Cognitive Radio Networks raises several open issues and security concerns. CRNs suffer from both classical wireless networks vulnerabilities and threats, and new threats related to their inherent functionalities. In this paper, an overview of the cognitive radio networks and their security challenges will be provided. Both, the traditional and new security threats that emerged from these promising networks are addressed. The paper will also focus on the Primary User Emulation (PUE) attack as one of the main specific attacks targeting CRNs and analyze some proposed countermeasure. Furthermore, CRN security requirements are introduced.

Within few years, vehicles communicating with other vehicles to provide information about road conditions, accidents, fires, or emergency cases, will be a reality. Vehicles will also have access to the internet. As a result, future vehicles networks security should be designed to cope with various kinds of attacks. Furthermore, security requirements including confidentiality, integrity, privacy, and nonrepudiation should be enforced. This paper introduces multi-level security architecture for vehicular ad hoc networks (VANETs). Based on this architecture, the security protocols for Vehicle-to-Vehicle (V2V), Vehicle-to-Roadside Unit (V2R), and Roadside-to-Roadside Unit (R2R) will be presented.

The Traditional Power utilities are gradually moving towards the Smart Grids. These Grids deploy a very large number of smart meters at the consumers' sites using bi-directional communication networks based on Internet protocols. With the reliance on these protocols, the Smart Grids become vulnerable to various cyber-attacks. Smart meters collect consumption data and allow customers other useful functions. Consumers are worried about the privacy, integrity, availability, and confidentiality when managing their future power consumption. In an attempt to contribute to the protection of these smart meters from attacks, two approaches based on cryptographic protocols are proposed for securing the direct connection of smart meters to collectors.