Stakeholders make various types of decisions with respect to requirements, design, management, and so on during the software development life cycle . These decisions are typically not well documented and classified due to limited human resources, time, and budget . In this paper, we aimed at automaticallyclassifying decisions into five types to help stakeholders better document and understand decisions . We experimented and evaluated 270 configurations of feature selection, feature extraction techniques, and machinelearning classifiers to seek the best configuration for classifying decisions . Our experiment results show that (1) feature selection candecently improve the classification results; (2) ensemble classifiers canoutperform base classifiers . (3) BoW + 50% features selected by feature selection with anensemble classifier . combines Na\”ive Bayes (NB), Logistic Regression (LR), and (SVM) achieves the best classification result (witha weighted precision of 0.750, a weighted recall of 0 .739, and a weighted recalls of . 0.727) among all the configurations

Author(s) : Liming Fu, Peng Liang, Xueying Li, Chen Yang

Links : PDF - Abstract

Code :

Keywords : decisions - feature - classification - classifiers - weighted -

Leave a Reply

Your email address will not be published. Required fields are marked *