Estimating Software Effort with Minimum Features Using Neural Functional Approximation

Abstract

The aim of this study is to improve software effort estimation by incorporating straightforward mathematical principles and artificial neural network technique. Our process consists of three major steps. The first step concerns data preparation from each considered database. The second step is to reduce the number of given features by considering only those relevant ones. The final step is to transform the problem of estimating software effort to the problems of classification and functional approximation by using a feedforward neural network. Experimental data are taken from well-known public domains. The results are systematically compared with related prior works using only a few features so obtained, yet demonstrate that the proposed model yields satisfactory estimation accuracy based on MMRE and PRED measures.

Impact of Product Complexity on Actual Effort in Software Developments: An Empirical Investigation

“…When it comes to setting a threshold as minimum confidence, we try to borrow ideas from the performance assessment for effort estimation models. When assessing effort estimation models, in general, both Percentage Relative Error Deviation within x (PRED(x)) and Mean Magnitude Relative Error (MMRE) adopted 25% as measurement threshold [11,13,15]. Similarly, here we define that an association rule is acceptable if its incorrect predictions are less than 25% of applicable instances in a dataset.…”

A differential evolution-based model to estimate the software services development effort

“…Dasharnais is one of the most common datasets in the field of software effort estimation . Although this dataset is relatively old, it has been widely employed in many of recent research studies . In this dataset, there are 81 projects related to a Canadian software company, out of which, four projects include missing values, and the remaining 77 projects are considered in the evaluation process.…”

Fine-Grained Work Element Standardization for Project Effort Estimation

“…Project managers (PMs) must decide what data are to be collected to suit the applicable domain. Jodpimai, Sophatsathit, and Lursinsap [15] explored the relationship of different project dimensions to select only a handful of relevant cost drivers as oppose to standard 16 factors in existing approaches, yet yielding similar outcome. The needs for standardizing its deliverables and development process are key factors to software products.…”

“…For example, the COCOMO model [3] uses 17 cost drivers in the estimation process. Jodpimai, Sophatsathit, and Lursinsap [15] found that only a handful of cost drivers were effective factors that could derive as accurate estimation as the comparative models without employing the full-fledge parameter set. Moreover, fewer cost drivers translated into faster computation time.…”

“…Features that result in low value will be less relevant and thus eliminated. Finally, only those features that are related to effort estimation in the form of a mathematical function will be retained [15].…”