Analytically selecting which tests are to be executed in a continuous delivery process

A method for analytically selecting which tests are to be executed in a continuous delivery process, the method comprising:

• constructing a code dependency tree to analyze tests affected by changes in code, wherein said code dependency tree illustrates which code is utilized by a particular test;
• analyzing, by a processor, said code dependency tree to determine which tests need to be executed based on whether those tests depend on a changed code;
• eliminating tests in said code dependency tree for execution that do not depend on said changed code;
• analyzing, by said processor, historical execution records for remaining tests from said code dependency tree that were not eliminated to obtain an execution rate and a failure rate for each of said remaining tests, wherein said execution rate indicates how often a test is being executed, wherein said failure rate indicates how often a test has failed;
• generating a score for each of said remaining tests from said code dependency tree that were not eliminated based on said execution rate and said failure rate;
• and
• executing those tests of said remaining tests from said code dependency tree that were not eliminated based on having a score exceeding a threshold.