Date of Graduation
Statler College of Engineering and Mineral Resources
Lane Department of Computer Science and Electrical Engineering
Optimization has been the goal of almost every human thought and action. With the growing computational capabilities, solutions to problems are also exponentially increasing. With rising demand for data and analytics on this big data, solutions to problems are also multiplied. Although, strong statistical and analytical reasoning has helped narrow down many solutions into set of feasible applicable solutions; Experts also seek knowledge of how these solutions have improved and what decisions are needed to be taken for remaining solutions to improve. This problem persists in software engineering with managers and domain experts taking decisions deciding future course of the project.;This thesis proposes a method for optimizing solutions along with providing decisions that help improve a solution. Literature supports that landscape visualization gives inside details on data behavior. A method is proposed which takes benefit of visualizing data and improving solutions based on their position in landscape. CrossTrees are built by grouping data based on their similarities and generating contrast sets between these groups, which help experts to gain insight from knowledge. This thesis tests CrossTrees with two models which simulate software projects: POM3 and XOMO. Models are simulated with a widely used genetic algorithm NSGA-II, to generate set of highly optimized solutions. When compared to NSGA-II, CrossTrees recommendations are more effective and faster to generate. Further as problem complexity increases, NSGA-II takes (worst-case) polynomial time while CrossTrees runs in linear time.
Lekkalapudi, Naveen Kumar, "Cross Trees: Visualizing Estimations using Decision Trees" (2014). Graduate Theses, Dissertations, and Problem Reports. 6055.