Abstract:
Numerous project management applications have made extensive use of machine learning
algorithms wherein risk assessment module plays an important role in project timeline
management. In order to increase the likelihood that software projects will be
successfully developed, research into project management applications that involve risk
assessment has become increasingly popular in recent years, especially with the use
of machine learning techniques to pinpoint project risk indicators before the project’s
development even begins.
In this research, we have applied five cutting-edge machine learning techniques namely
Artificial Neural Network (ANN), Logistic Regression (LR), Naive Bayes (NB), Random
Forest (RF), and K-Nearest Neighbors (K-NN) on two different risk datasets. Moreover,
hyper-parameter tuning is applied using different parameters in each machine learning
technique and find the optimal risk predictions in each of them. On the basis of comprehensive
literature review, a methodology was proposed which made it easier for the
project manager and subject matter experts to plan and reduce risks at an early stage by
determining the level of risk involved. Moreover, different experiments are performed on
benchmark datasets to compare the models performance based on the applied parameters.
This research will assist domain experts, data analysts, developers, and researchers
to better develop ML models and make effective predictions which provide guidance towards
making effective decisions to meet the project’s deadline and organizational goals.
