Using machine learning methods to classify tasks by priority in IT-projects
DOI:
https://doi.org/10.31449/upinf.240Keywords:
IT-project management, task prioritisation, machine learning, multiclass classification, data imbalanceAbstract
Prioritisation and prioritising tasks remains a challenge in effective project management. There are many approaches to prioritisation, such as MoSCoW, binary search tree and others. However, all these techniques are labour intensive, subjective and inflexible. In this paper, we discuss machine learning based approaches for automatic prioritization. The main goal of our work is to investigate how machine learning techniques can be used to help project managers prioritize tasks more efficiently in IT-projects. To this end, we developed a classification model for automatic prioritization on a set of 600000 records following the CRISP-DM step-by-step process. Most tasks in IT-projects are labelled with the highest priority, which poses a challenge for modelling as well as for the efficiency of project execution. We show that it makes sense to classify tasks into fewer priority groups, which in turn contributes to the accuracy of the classification model.