Master of Science (MS), Ohio University, 2007, Electrical Engineering & Computer Science (Engineering and Technology)

The intent of this thesis is to improve on existing algorithms for determining classification rules by reducing the computational time to generate the reducts of an information system. Determining all reducts is an NP (Non-deterministic Polynomial time) complete problem and, therefore, as the data set grows in size, the time required for computation rapidly exceeds what is practical. This thesis has been able to significantly reduce the amount of time it takes to perform these computations. While the problem is still NP complete, the amount of time required by the methods introduced is less than other well-known methods provided by other software packages such as Rosetta [Ohr99] and RSES [RSES2]. Despite the reduct generation time improvements, larger databases still take far too long for effective reduct determination; therefore, heuristic non-exhaustive methods were also evaluated. In practical applications of rough sets, it is important that the obtained reducts retain most of the information about the original problem. In these applications, reducts of a dataset are used as classifiers to determine the “rules” for classification. The second half of this thesis proposes a method for rapidly producing effective classifiers of sufficient quality to get classification results of equal or better quality compared to exhaustive methods. The proposed method gives results that are at, or near, the same quality as those obtained from using the exhaustive method in only a fraction of the computational time.

Committee: Janusz Starzyk (Advisor) Subjects:

Master of Science, University of Akron, 2009, Computer Science

Reverse engineering malware has been an integral part of the world of security. At best it has been employed for signature logging malware until now. Since the evolution of new age technologies, this is now being researched as a robust methodology which can lead to more reactive and proactive solutions to the modern security threats that are growing stronger and more sophisticated. This research in its entirety has been an attempt to understand the in and outs of reverse engineering pertaining to malware analysis, with an eye to the future trends in security.Reverse engineering of malware was done with Nugache P2P malware as the target showing that signature based malware identification is ineffective. Developing a proactive approach to quickly identifying malware was the objective that guided this research work. Innovative malware analysis techniques with data mining and rough sets methodologies have been employed in this research work in the quest of a proactive and feasible security solution.

Committee: Kathy J. Liszka PhD (Advisor) Subjects: Computer Science; Engineering; Experiments; Systems Design