Software Effort Estimation using Machine Learning Techniques

Abstract

Success of the software development companies is mostly dependent on the best effort prediction. If the predicted
effort is somewhat correct, then the company can find relief from the great tension of hurrying up the employees to get
the job done within targeted time. There are many estimation methods, techniques and tools that are available. But it
is very difficult to select the best one for a particular project. Each method has its own advantages and disadvantages.
And also the effort estimation depends on various parameters. It is the responsibility of the project manager to select the
best tool for his project. Based on the historical data, the project manager can find effort value of the new project after
applying some statistical methods and data mining techniques on that data. The main aim of this work is to reveal how
much accurate are data mining-classification techniques on software project effort prediction datasets when we perform
analogy based effort estimation.