期刊名称：International Journal of Advances in Engineering and Management
电子版ISSN：2395-5252
出版年度：2021
卷号：3
期号：7
页码：1509-1517
DOI：10.35629/5252-030713481353
语种：English
出版社：IJAEM JOURNAL
摘要：The requirement elicitation phase of software development is an important process in the software development life cycle (SDLC) as it entails an extensive and a thorough finding of information process gathered from all team members associated with the built of software. This phase of the software development process is associated with uncertainties and risks such that if the risks are visualised early enough, managing them will improve software quality and enhance timely deliverables of deadlines. Data visualisation is concerned with the task of putting data into a chart, graph or other visual formats that help inform analysis and interpretation. The widely used data visual formats are; frequency tables, crosstabulation tables, bar charts, line graphs, pie charts, Heat maps, scatter graphs and histograms. The histogram data visual format has been considered and used in this paper to interpret the risk levels, their impacts and its probability across software projects for informed decision making by project managers and developers. The dataset for software risk prediction was adopted to visualise the risk levels and their impacts as well as their probability of occurrence in the software projects. This was necessitated to provide a pictorial view for risk decision support system and to educate the software developers on what to look out for as to foster better planning and timely project deliveries. A machine learning technique was applied to predict the software risk levels and their levels of impact which was represented using histograms in R. The Architecture of the Naïve Bayes Risk prediction system was also developed and an analysis of the performance metrics of the risk attributes with respect to the risk levels across the projects was also done using R. The Naïve Bayes Risk Prediction and Analysis Model (NBRPAM) had a plot of the performance matrix of the variables which showed that probability and priority are the most significant variables in predicting Risk levels in software projects.
关键词：Requirement Elicitation;Data Visualisation;Risk Levels;Software Development Life Cycle (SDLC);Histogram;Naïve Bayes;R programming
