Software Defects Prediction based on ANN and Fuzzy logic using Software Metrics

Software Quality analysis is one of the significant criteria required to explore the software life and additionally software reliability. Software quality is been characterized under various parameters. Software risk analysis is one such basis required to distinguish the software reliability. At the point when software is arranged or being created by the sort of software and in addition the endeavors required to build up the software by and large characterizes the software hazard. For example, the accessibility of the required software, equipment, man power all are the prescient hazard factors. In this work, these all hazard factors are characterized under the fuzzy outing the demonstration. In light of this fuzzy estimation to the some weightage is been allotted to these all hazard factors. However conventional metrics approaches, numerous predictable methodologies are inadequate in this regard as well as on a very basic level conflicting. Other than this the paper additionally legitimizes Neural Networks as a superior contrasting option to formal techniques in initiating times of software improvement lifecycle. A fuzzy logic reputable paradigm is proposed for predicting software defect density on individual phases of the SDLC. The perceptive precision of the proposed model is applicable utilizing five real software project data, Approval comes about are appealing and Measures in view of the MMRE and BMMRE fundamentally as the software project estimate growths.