A Linear Regression Model with Exponential Transformation for Software Effort Estimation

ABSTRACT—Software Engineering plays a vital role in software development process. Software effort Estimation usually takes place in early stage of software life cycle. Estimation relies upon previous experience. Software development based on accurate estimates because badly chosen estimate causes problems during the performance of software development process. A linear Regression model with exponential transformation is proposed to evaluate the estimation of software effort from use case diagrams. This model can be used in early stages of software life cycle to improve the software effort estimation. Linear Regression is used to find out the relationship between variables to get accurate results in software effort. For accurate results in regression, the data obtained in the effort should be normally distributed. For Normal Distribution exponential transformation of data is applied. The exponential function is used to model the relationship in which a constant change in the independent variable gives the same proportional change in the dependent variable. Linear Regression is applied on normalized data to improve the accuracy of software effort estimation.