Fuzzy-Provenance Architecture for Effort Metric Data Quality Assessment

Software companies rely on stored metric data in order to track and manage their projects, through analyzing, monitoring and estimating software metrics. If managers cannot believe the metrics data, the product that is being developed is fated to fail. Currently, the assessment of software effort is subjective and derived mainly through managers’ assumptions, which is fundamentally an error-prone process. We present an architecture for assessing data quality of software effort metric based on data provenance associated with a mechanism of logical inference (fuzzy logic). The contribution is to provide an assessment to search evident reasons for a low quality in order to ensure that the metrics can be used with sufficient reliability.