Simple empirical software effort estimation model

Abstract

Context: An effort estimation model with more than 20 parameters is not very useful at early conceptual phase if you don't have a logical approach for specifying the input values. Goal: This paper presents a simple approach for predicting software development effort. Method: The regression model uses product size and application types to predict effort. Product size is measured in terms of the equivalent source lines of code. The analysis is based on empirical data collected from 317 very recent projects implemented within the United States Department of Defense over the course of nine years beginning in 2004. Result: Statistical results showed that source lines of code and application type are significant contributors to development effort. Conclusion: The equation is simpler and more viable to use for early estimates than traditional parametric cost models. The effect of product size on software effort shall be interpreted along with application domain.