Data Quality Model-based Testing of Information Systems

Abstract

This paper proposes a model-based testing approach by offering to use the data quality model (DQ-model) instead of the program's control flow graph as a testing model. The DQ-model contains definitions and conditions for data objects to consider the data object as correct. The study proposes to automatically generate a complete test set (CTS) using a DQmodel that allows all data quality conditions to be tested, resulting in a full coverage of DQ-model. In addition, the possibility to check the conformity of the data to be entered and already stored in the database is ensured. The proposed alternative approach changes the testing process: (1) CTS can be generated prior to software development; (2) CTS contains not only input data, but also database content required for complete testing of the system; (3) CTS generation from DQ-model provides values against which the system can be further tested. If the test results correspond to the values obtained during CTS generation, the system under test shall be considered to have been tested according to DQ-model. Otherwise, the user can verify the cause of the differences that may occur due incorrect software, as well as an inaccurate specification.