Loading...
Please wait, while we are loading the content...
A Collective Study of PCA and Neural Network based on COCOMO for Software Cost Estimation
| Content Provider | Semantic Scholar |
|---|---|
| Author | Waghmode, Rina M. Patil, Lalit V. Joshi, S. D. |
| Copyright Year | 2013 |
| Abstract | Estimating cost is a very wearisome activity in all aspect. A person with broad scope and good thinking for the future makes more precise decisions. It helps in governing and planning the software risks which are admirably correct and precise. In 1960 regression analysis and mathematical formulae were practiced to determine cost. We need to think more than simply putting numbers into a formula and accept the results to attaining the accuracy of software cost estimation. The changing methods of estimating software cost have made the researchers to think diversely. Barry Bohem birthed COCOMO model for software cost estimation in 1981 which is considered to be more efficient as compared to previous models. Thereafter number of researchers has been trying to improve the efficiency by keeping the base of COCOMO model. The paper drafts a novel variable reduction technique called feed-forward neural network with PCA to measure the estimation model accuracy. This is based on a COCOMO sample data set which collects and maintains a large software project data repository. PCA is a kind of classification method which can reduces number of factors into a few absolute factors. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://research.ijcaonline.org/volume74/number16/pxc3890099.pdf |
| Alternate Webpage(s) | http://www.ijcaonline.org/archives/volume74/number16/12970-0099?format=pdf |
| Language | English |
| Access Restriction | Open |
| Subject Keyword | Artificial neural network Biological Neural Networks COCOMO Estimated Feedforward neural network Mathematics Research data archiving Software development effort estimation Software project management Statistical classification |
| Content Type | Text |
| Resource Type | Article |