Variable importance and interaction measures are crucial to breaking open the "black box” of machine-learned classifiers. The existing metrics, however, are data-driven and lack a solid mathematical foundation, resulting in misleading conclusions on certain types of data. We propose feature power: a new variable importance measure based on the Shapley value of cooperative game theory. We evaluate the validity of this new measure and the behavior of feature power in comparison to existing variable importance metrics.
Time is extremely important to us nowadays because of the fast-paced life and society. However, people still spend substantial time waiting in lines at restaurants. Although some hot applications, for example, Yelp and OpenTable, allow people to make a reservation with restaurants, customers must arrive at the restaurant to sign on a waitlist if the restaurant is full. In that case, a customer has no idea of how many parties ahead of him or how long he needs to wait until he actually arrives.
Twitter spam has been a challenging and critical problem. Previously, researchers have proposed many different machine learning based methods to address this problem. One limitation of those existing solutions is that there is often not enough labeled data. Another issue is feature selection. There is no universal standard on feature selection for spam detection. Researchers often manually choose features based on individual experience and observation.
Security is a growing concern in the software industry. Every year the industry takes billions of dollars in losses due to software security breaches. These breaches are carried out in the form of different attacks, such reverse engineering attacks, copyright violations, piracy, hacking, code tampering, etc. In this thesis, we examine using code obfuscation for protection against reverse engineering attacks. Code obfuscation is also used for anti-tampering, anti-piracy or watermarking.
Anti-cancer drug development requires an understanding of the mechanism of action (MOA). Histological analysis of images taken from xenografts enable characterization of how the tissue reacts to different compounds. Immunohistochemical (IHC) stains highlight a specific aspect of the tissue. Registering a two-dimensional (2D) IHC image with its corresponding 2D cellular morphology image facilitates the study of complex immunological behavior. This thesis describes the design and implementation of a multi-stage process to register histological images.