Enhance Rule Based Detection for Software Fault Prone Modules

摘要: Abstract Software quality assurance is necessary to increase the level of confidence in developed software and reduce overall cost for developing projects. The problem addressed this research prediction fault prone modules using data mining techniques. Predicting allows managers allocate more testing resources such modules. This can also imply a good investment better design future systems avoid building error models that are based upon from previous projects identify fault-prone current similar development project, once similarity between established. In paper, we applied different rule-based classification techniques on several publicly available datasets NASA repository (e.g. PC1, PC2, etc). goal was classify into either or not paper proposed modification RIDOR algorithm which results show enhanced than other terms number extracted rules accuracy. implemented learns defect static code attributes. Those attributes then used present new predictor with high accuracy low rate.