INTELLIGENT TEST SELECTION BASED ON BIG-DATA ANALYSIS AND MACHINE LEARNING

Smart overview of the Invention

The patent application outlines a system and method for intelligently selecting tests to run on modified source code of enterprise applications. It leverages machine-learning algorithms to predict which tests are necessary based on changes made to the application code. This approach aims to streamline the testing process in a continuous integration environment, where code is frequently updated by multiple developers.

Background

Enterprise applications, often composed of millions of lines of code, require rigorous testing whenever modifications occur. Traditionally, test engineers manually select test cases based on experience, which can be challenging and unreliable as the number of test cases increases. The manual selection process is time-consuming and can miss potential errors, highlighting the need for a more efficient and systematic approach.

Machine Learning Integration

The proposed system uses machine-learning models to predict an initial set of test cases for modified code. This prediction is based on the type of modifications made, such as changes to existing functions or the addition of new features. The test prediction system collects data on the test engineer's adjustments to the initial set, such as adding or removing tests, and uses this data to refine the model over time, reaching a big-data scale.

Adaptive Testing Process

The system allows test engineers to modify the suggested test suite by adding or removing test cases, enhancing the robustness of the selection. These modifications are tracked and used to continuously improve the machine-learning model's predictive capabilities. The system can execute automated scripts for the finalized test suite, optimizing the testing process and reducing resource burdens.

Implementation and Benefits

A computer-implemented method is described, involving the detection of code modifications, determination of relevant tests, and collection of test selection data. This data trains the machine-learning model, which predicts test sets for future code changes. The system offers a more reliable and efficient testing process, reducing costs and improving the accuracy of error detection in enterprise applications.