Developing an Intelligent Generator for Semi-Actual Test Data
The actual test data generation is one of the difficult and expensive parts of applying software-testing techniques. Many of the current test data generators suffer from the reduction of user’s confidence in generated test data and testing process. This is because of focusing on developer and database administrator viewpoints regardless of users concerns and focusing on data type and structure regardless of meaning. This paper proposes a model of an intelligent generator for semi-actual test data with the aim of increasing users confidence in software testing. The model uses samples of real data as a resource data and a set of efficient generation techniques based on statistical methods such as permutations, combination, sampling, and statistical distributions. The selection of the suitable structure and generation technique is based on one of the intelligent soft computing techniques such as fuzzy logic, neural network, heuristic, or genetic algorithm. The generated test data is validated according to the data specifications then tested by one of the normality testing techniques to be close to the real world or environment of the testing processes. This model offers the ability of simulating real environments.
Key Words: Software Testing, Test Data Generation, Semi-Actual Data, Intelligent Generator, Simulation.
The IIUM Engineering journal follows the open access policy.