"Polymorphic malware is a menace to modern computing and a strain on business productivity. The challenge faced by antivirus technology is that there is not enough time for new variants of this type of malware to be collected, sent to antivirus companies, and analyzed, and for signatures to be created and returned to customers. To attempt to address this problem, we explore the classification of malware using machine learning. We compare some classifiers for malware and present a carefully selected set of attributes that result in good classification between malware and clean programs. We discuss the application of this research to security technologies. We apply well-known machine-learning algorithms to help address the problem of malware classification. From experience, I know that the nature of antivirus research is reactive. Researchers have to and tend to focus on technical problems or addressing particular families of malware. This research was done for my master's thesis at UC Irvine where there was less pressure on me to focus on the daily problems of antivirus research. I approach the general problem of polymorphic malware classification instead of the topical problems of detecting particular malware. In addition, this research uses machine-learning techniques, which are seemingly underutilized by industry to solve security problems but that are used by other computing disciplines with success. Hence, this is research that is novel in its scope and in its techniques. In our best results, we achieved a 98.56% classification rate for malware using only seven executable file-format features: DebugSize, ImageVersion, IatRVA, ExportSize, ResourceSize, VirtualSize2, and NumberOfSections. Industry and other security researchers will benefit from receiving this research, the whitepaper, and the tool."