THE ARTIFICIAL NEURAL NETWORK BASED HIERARCHICAL PROBABILISTIC APPROACH FOR RISK ASSESSMENTS OF IT PROJECTS

This paper aims to attain high accuracy of risk assessment during monitoring and evaluation of software risk factors by reconstruction of the pattern for risk recognition and estimation, the research helped construct a functional model for risk assessment. Computer programmes have numerous features that also maximize the risks for failure of the project. The current work thereby incorporates the Artificial Neural Network (ANN) based Hierarchical Probabilistic Approach (HPA) for an early stage approach to the software risk assessment, in order to solve this problem. The questionnaire was designed to classify the risk model base as well as provide adequate data for the suggested model. This brought forward a model for predictions using historical data and taking into account other important common risk factors. The research showed that the employed method can be utilized effectively for the identification of risk effects in the entire phases of software development project. We categorized the top risks into human skills, knowledge, the experience of the staff level and difficulties in applying software management and organization presented as software cost risks in the total risk evaluation. Artificial Neural Network ANN represents one of the promising modeling techniques for multi objective software risks such as: cost, scheduling, quality and other business risks.