Font Size: a A A

The Research And Improvement Of Capture-recapture Models In Software Engineering

Posted on:2017-04-19Degree:MasterType:Thesis
Country:ChinaCandidate:Q ShanFull Text:PDF
GTID:2308330485458775Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Since capture-recapture model was introduced into the software engineering research community in 1972 by Mills, capture-recapture model has caught much attention and adoption as an outstanding estimation model. The thesis briefly introduced the background knowledge, research status and model classification of capture-recapture model, as well as the research results and limitations. On this basis, the thesis specified an improved research method based on the virtual team and expounded the process and advantages of this research method.This research method requires plenty of computing resource and high computational efficiency. Since existing tools are not able to satisfy such demand, we developed an capture-recapture calculator to help carry out researches. The thesis discussed the disadvantages of existing tools and the requirements, design and implementation of the capture-recapture calculator.For the purpose of better understanding of the improved research method, the thesis introduced an experiment in detail. The research purpose of this experiment was to generate comprehensive assessment of estimators and the influence between coefficient of covariation and estimators. Five estimators were adopted into this experiment:Mo-MLE, Mh-Chao, Mh-JK, Mt-Chao and Mt-MLE. With the assistance of the calculator, we constructed virtual teams and calculated the results. According to the experiment results and rigorous statistical tests, we could draw two conclusions. The performance of Mh-JK is the best among five estimators. The combination of MrMLE and coefficient of covariation could also gain precise estimated results.
Keywords/Search Tags:Capture-recapture model, Estimators, Virtual team, Coefficient of covariation
PDF Full Text Request
Related items