DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Lee, Jae-Kyu | - |
dc.contributor.advisor | 이재규 | - |
dc.contributor.author | Jun, Eung-Sup | - |
dc.contributor.author | 전응섭 | - |
dc.date.accessioned | 2011-12-27T04:19:01Z | - |
dc.date.available | 2011-12-27T04:19:01Z | - |
dc.date.issued | 2001 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=169455&flag=dissertation | - |
dc.identifier.uri | http://hdl.handle.net/10203/53361 | - |
dc.description | 학위논문(박사) - 한국과학기술원 : 경영공학전공, 2001.8, [ ix, 138 p. ] | - |
dc.description.abstract | A number of software effort estimations have been attempted using statistical models, case-based reasoning, and neural networks. The research results showed that the neural network models perform at least as good as the other approaches. However, since the computing environment changes so rapidly in terms of programming languages, development tools, and methodologies, it is very difficult to maintain the performance of estimation models for the new breed of projects. So we propose using the relevant cases for a neural network model, whose cost is the decreased number of cases. To balance the relevance and data availability, the qualitative input factors are used as the criteria of data classification. With the data sets that have the same values for certain qualitative input factors, we can eliminate the factors from the model making the reduced neural network models. So we need to seek the optimally reduced neural network model among them. To find the best reduced model heuristically, we propose the ADD and DROP algorithms that add and drop individual qualitative factors one at a time in selecting the data set and associated reduced neural network model. According to our experiment with the factors that have been adopted in COCOMO model, the mean error rate was significantly reduced to 18.6%. In order to increase the estimation performance, we design the neural network model and suggest a case-based approach with optimally similar cases selected by heuristic search algorithms. We could find the efficiency and effectiveness in reducing estimate errors and the computation time outstandingly by the empirical test. Also, we are required to redefine the variables that are used for the existing software effort estimation model compatible with the suggested model. In this research, we define the input factors and design the neural network architecture of our software effort estimation model. We especially focus the research issues on finding the optimal cases for... | eng |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | Similarity meausre | - |
dc.subject | Neural network model | - |
dc.subject | Similar cases | - |
dc.subject | Software efort etimation | - |
dc.subject | Reduced neural network model | - |
dc.subject | 축약형 신경망모델 | - |
dc.subject | 유사도측정 | - |
dc.subject | 인공신경망모델 | - |
dc.subject | 최적사례선별 | - |
dc.subject | 소프트웨어공수예측 | - |
dc.title | Software effort estimation systems by neural network using optimally similar cases | - |
dc.title.alternative | 최적사례 선별과 인공신경망 모델을 이용한 소프트웨어 공수예측 시스템 | - |
dc.type | Thesis(Ph.D) | - |
dc.identifier.CNRN | 169455/325007 | - |
dc.description.department | 한국과학기술원 : 경영공학전공, | - |
dc.identifier.uid | 000929025 | - |
dc.contributor.localauthor | Lee, Jae-Kyu | - |
dc.contributor.localauthor | 이재규 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.