Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/9090
Title: AI Models for Software estimation
Researcher: Deshpande, Manojkumar
Guide(s): Bhirud, Sunil
Keywords: AI Models
Computer Sciences
Upload Date: 23-May-2013
University: Narsee Monjee Institute of Management Studies
Completed Date: 04/08/2012
Abstract: The software effort and cost estimation is essential to provide key information about project planning and control. An accuracy of estimates is one of the most important factors for projectsand#8223; success. In the research literature, number of estimation models, methods, techniques and tools have been proposed. However, their accuracy of estimates can be challenged for complex real world projects. This thesis systematically reviews different software effort estimation model, methods, technologies and tools. Theoretical analysis and empirical experiments show that the persistent problems in software effort estimation is not only technical one, but also fundamental theoretical, where new theories and mathematical models still need to be sought such as what dominates development effort in a large group ? How labor is traded with time in a project ? What is the optimal organization form of a large software project ?. The survey has been conducted, where software project managers were asked to respond to the questionnaire. This survey has provided important input to understand pain areas and decide research focus. This thesis also presents comparative analysis of various estimation models and underlying AI techniques and technologies. The assumption is made that software project life cycle is divided in early stage, intermediate stage and final stage. At early stage, Request for Proposal and Software Requirement Specification documents are available. Similar project identification technique is presented which involved LSA and ontology for finding semantically similar documents. The AI model has been proposed to mimic human expert estimator, who applies the experience and estimates by referring textual information of project requirements. During Intermediate stage, requirements are clear as well as analysis and design models are available. COCOMO II in combination with AI technologies such as neural networks, fuzzy logic and neuro-fuzzy inferencing system are evaluated.
Pagination: 129p.
URI: http://hdl.handle.net/10603/9090
Appears in Departments:Department of Computer Engineering

Files in This Item:
File Description SizeFormat 
01_title page.pdfAttached File312.81 kBAdobe PDFView/Open
02_abstract.pdf85.72 kBAdobe PDFView/Open
03_acknowledgements.pdf84.61 kBAdobe PDFView/Open
04_table of content.pdf199.52 kBAdobe PDFView/Open
05_chapter 1.pdf170.16 kBAdobe PDFView/Open
06_chapter 2.pdf314.15 kBAdobe PDFView/Open
07_chapter 3.pdf1.21 MBAdobe PDFView/Open
08_chapter 4.pdf398.13 kBAdobe PDFView/Open
09_chapter 5.pdf88.15 kBAdobe PDFView/Open
10_appendix a-b.pdf472.74 kBAdobe PDFView/Open
11_references.pdf2.82 MBAdobe PDFView/Open


Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.

Altmetric Badge: