Machine Learning Approach for Cost and Effort Estimation in Agile Development Process

Software project management is a key area in the field of computer science as software

now-a-days impacts every area related to human life. Managing software means the process for

development and the maintenance of software must be completely controlled using various

pre-defined set of rules. As the software development process has to follow various

parameters and a well-defined life cycle to ultimately deliver all the requirements gathered from

the customers hence it has become quite time consuming and expensive process. It is also an evident

fact that failure in software is caused mainly due to faulty practices used in

project management. Using the right and optimised practices for software management helps

both client as well as developers. Because of all the factors the need for highly reliable

software is increasing. The reliability of software is mainly dependent on two factors:

the selection of proper model for development and the estimation of various parameters.

During the last few decades, the former area has been a research interest for many researchers

resulting in development of many reliability models. Hence, currently parameter estimation is

considered to be a primary activity in software reliability prediction and broadly the

most important aspect of software project management. Software reliability models only become

useful if they provide a correct and optimal estimation of various parameters.

A successfully completed project means that the project is developed within the planned budget and

timeline which is mostly related to accurate effort and cost estimation whereas inaccurate

estimation of effort and cost results in failure of a project in context of delivery

time, cost and other parameters. Hence the most important parameters requiring accurate

estimate in terms of software projects are effort and cost. The accuracy of the estimation of

these two vital parameters depends on the correct estimation of size of the project to be

developed, and the ability to convert the size estimate into man hours, duration and cost.