A novel fuzzy mechanism for risk assessment in software projects

Risk management is a vital factor for ensuring better quality software development processes. Moreover, risks are the events that could adversely affect the organization activities or the development of projects. Effective prioritization of software project risks play a significant role in determining whether the project will be successful in terms of performance characteristics or not. In this work, we develop a new hybrid fuzzy-based machine learning mechanism for performing risk assessment in software projects. This newly developed hybridized risk assessment scheme can be used to determine and rank the significant software project risks that support the decision making during the software project lifecycle. For better assessment of the software project risks, we have incorporated fuzzy decision making trial and evaluation laboratory, adaptive neuro-fuzzy inference system-based multi-criteria decision making (ANFIS MCDM) and intuitionistic fuzzy-based TODIM (IF-TODIM) approaches. More significantly, for the newly introduced ANFIS MCDM approach, the parameters of ANFIS are adjusted using a traditional crow search algorithm (CSA) which applies only a reasonable as well as small changes in variables. The main activity of CSA in ANFIS is to find the best parameter to achieve most accurate software risk estimate. Experimental validation was conducted on NASA 93 dataset having 93 software project values. The result of this method exhibits a vivid picture that provides software risk factors that are key determinant for achievement of the project performance. Experimental outcomes reveal that our proposed integrated fuzzy approaches can exhibit better and accurate performance in the assessment of software project risks compared to other existing approaches.