Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.581905
Title: Pragmatic cost estimation for web applications
Author: Letchmunan, Sukumar
Awarding Body: University of Strathclyde
Current Institution: University of Strathclyde
Date of Award: 2012
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
Cost estimation for web applications is an interesting and difficult challenge for researchers and industrial practitioners. It is a particularly valuable area of ongoing commercial research. Attaining on accurate cost estimation for web applications is an essential element in being able to provide competitive bids and remaining successful in the market. The development of prediction techniques over thirty years ago has contributed to several different strategies. Unfortunately there is no collective evidence to give substantial advice or guidance for industrial practitioners. Therefore to address this problem, this thesis shows the way by investigating the characteristics of the dataset by combining the literature review and industrial survey findings. The results of the systematic literature review, industrial survey and an initial investigation, have led to an understanding that dataset characteristics may influence the cost estimation prediction techniques. From this, an inv estigation was carried out on dataset characteristics. However, in the attempt to structure the characteristics of dataset it was found not to be practical or easy to get a defined structure of dataset characteristics to use as a basis for prediction model selection. Therefore the thesis develops a pragmatic cost estimation strategy based on collected advice and general sound practice in cost estimation. The strategy is composed of the following five steps: test whether the predictions are better than the means of the dataset; test the predictions using accuracy measures such as MMRE, Pred and MAE knowing their strengths and weaknesses; investigate the prediction models formed to see if they are sensible and reasonable model; perform significance testing on the predictions; and get the effect size to establish preference relations of prediction models. The results from this pragmatic cost estimation strategy give not only advice on several techniques to choose from, but also give reliable results. Practitioners can be more confident about the estimation that is given by following this pragmatic cost estimation strategy. It can be concluded that the practitioners should focus on the best strategy to apply in cost estimation rather than focusing on the best techniques. Therefore, this pragmatic cost estimation strategy could help researchers and practitioners to get reliable results. The improvement and replication of this strategy over time will produce much more useful and trusted results.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.581905  DOI: Not available
Share: