USING STORY POINTS AND DECISION TREE TECHNIQUES TO ESTIMATE WORK AND COST IN AGILE SOFTWARE DEVELOPMENT

Abstract

For an Information Technology (IT) project to be resource-efficiently planned, early effort estimate is crucial. On the other hand, not much study has been done on the subject of artificial intelligence-based effort estimate in software development that is agile. With the use of learning-oriented methodologies, this research project expands the use of hybrid models made up of algorithmic models as a project-level effort estimating methodology in agile frameworks. The story point technique is used in agile approaches like Scrum to estimate effort, which is an arithmetic calculation of the manpower needed to finish a system release. This study used labelled historical data to predict the project's total cost in Pakistani rupees (PKR) and its completion time in days. using AdaBoost, random forests, and decision trees to increase prediction accuracy. Ten-fold cross-validation was used to train the models, and the relative error was employed to compare the outcomes with those found in the literature. The best accuracy is obtained by the three approaches combined bootstrap aggregation (bagging) ensemble, together with project categorization, improves the results even further.