Matching Items (3)
Filtering by
- All Subjects: Machine Learning
Description
The complexity of the systems that software engineers build has continuously grown since the inception of the field. What has not changed is the engineers' mental capacity to operate on about seven distinct pieces of information at a time. The widespread use of UML has led to more abstract software design activities, however the same cannot be said for reverse engineering activities. The introduction of abstraction to reverse engineering will allow the engineer to move farther away from the details of the system, increasing his ability to see the role that domain level concepts play in the system. In this thesis, we present a technique that facilitates filtering of classes from existing systems at the source level based on their relationship to concepts in the domain via a classification method using machine learning. We showed that concepts can be identified using a machine learning classifier based on source level metrics. We developed an Eclipse plugin to assist with the process of manually classifying Java source code, and collecting metrics and classifications into a standard file format. We developed an Eclipse plugin to act as a concept identifier that visually indicates a class as a domain concept or not. We minimized the size of training sets to ensure a useful approach in practice. This allowed us to determine that a training set of 7:5 to 10% is nearly as effective as a training set representing 50% of the system. We showed that random selection is the most consistent and effective means of selecting a training set. We found that KNN is the most consistent performer among the learning algorithms tested. We determined the optimal feature set for this classification problem. We discussed two possible structures besides a one to one mapping of domain knowledge to implementation. We showed that classes representing more than one concept are simply concepts at differing levels of abstraction. We also discussed composite concepts representing a domain concept implemented by more than one class. We showed that these composite concepts are difficult to detect because the problem is NP-complete.
ContributorsCarey, Maurice (Author) / Colbourn, Charles (Thesis advisor) / Collofello, James (Thesis advisor) / Davulcu, Hasan (Committee member) / Sarjoughian, Hessam S. (Committee member) / Ye, Jieping (Committee member) / Arizona State University (Publisher)
Created2013
Description
A new honors class created at Arizona State University utilizes a new "thinking" paradigm. The new paradigm is a problem solution using deductive logic and natural laws to replace the traditional acquisition and usage of detailed knowledge. When utilizing deductive logic, less time is required for students to learn, and students are able to resolve unique issues with minimal amounts of information. Students use their logic and processing skills to replace the traditional need of collecting large amounts of detailed information. The concepts taught in the class have come from the industry success of the Best Value (BV) approach developed by a leading research group at Arizona State University over the last 17 years. The research group identified the source of the industry's problem is due to the traditional business approach of management, direction and control (MDC). With over 1500 tests conducted, delivering $5.7B of services, with results showing: 30% decrease in cost, 30% increase in value, and customer satisfaction improvement by up to 140%, the Best Value (BV) approach has been identified as more efficient and can deliver better quality services than the traditional MDC approach. Through the research group's implementation of the new paradigm in higher education, the author identified a windfall effect that was able to give students understanding and an increased ability to cope with stressful situations, disease and extraordinary complications. It also exposed students to potentially harmful practices in their lives and has helped them to change. The study tested in K-12 proved potential value in exposing the paradigm to K-12 students, and what impact it may have on future professionals. The author's results include satisfaction rating of 9.5 (out of 10), increased career alignment by up to 113%, increased understanding of self by up to 70%, and a reduction of stress by up to 71%. The author's K-12 case studies aligned with the successful results shown in the industry and college classes run by the leading research group. The pattern of the new paradigm shows as resistance to it decreases, productivity, efficiency, processing speed, understanding, and effectiveness all increase.
ContributorsRivera, Alfredo (Author) / Kashiwagi, Dean (Thesis director) / Collofello, James (Committee member) / Nelson, Margaret (Committee member) / Barrett, The Honors College (Contributor) / Department of Management (Contributor) / Del E. Webb Construction (Contributor)
Created2013-12
Description
Machine learning tutorials often employ an application and runtime specific solution for a given problem in which users are expected to have a broad understanding of data analysis and software programming. This thesis focuses on designing and implementing a new, hands-on approach to teaching machine learning by streamlining the process of generating Inertial Movement Unit (IMU) data from multirotor flight sessions, training a linear classifier, and applying said classifier to solve Multi-rotor Activity Recognition (MAR) problems in an online lab setting. MAR labs leverage cloud computing and data storage technologies to host a versatile environment capable of logging, orchestrating, and visualizing the solution for an MAR problem through a user interface. MAR labs extends Arizona State University’s Visual IoT/Robotics Programming Language Environment (VIPLE) as a control platform for multi-rotors used in data collection. VIPLE is a platform developed for teaching computational thinking, visual programming, Internet of Things (IoT) and robotics application development. As a part of this education platform, this work also develops a 3D simulator capable of simulating the programmable behaviors of a robot within a maze environment and builds a physical quadrotor for use in MAR lab experiments.
ContributorsDe La Rosa, Matthew Lee (Author) / Chen, Yinong (Thesis advisor) / Collofello, James (Committee member) / Huang, Dijiang (Committee member) / Arizona State University (Publisher)
Created2018