Classification of Risk in Software Development Projects using Support Vector Machine

Traditionally, the lack of confidence in the system life cycle is expressed using the concept of risk. Nowadays, software development projects face various risks. However, the estimation and classification of risk, increased estimation of accuracy and reduced of uncertainty ultimately improve project outcomes. Therefore, in this paper, a Support Vector Machine (SVM) is used to model risk classification in software development projects. The proposed algorithm is compared with other methods in the literature such as Self Organizing Map (SOM) and K-Means based on measures of Classification Accuracy Rate (CAR) and Area Under Curve (AUC). According to the results, the proposed method exhibits superior CAR and AUC.