“Test Recommendation Engine”
summary
A probabilistic test-selection pipeline that ingests code-coverage (via OpenClover or JVM instrumentation), build meta and historical execution data into a DB, trains a recommendation model and predicts which regression tests to run for a code change. Uses static change-impact plus failure-history and collaborative filtering, trains a gradient boosted decision tree, retrains on new results, and reports a claimed 95% accuracy. Notes scalability issues on very large codebases and describes a predict–execute–retrain workflow.
AI-written summary · about
chapters (5)
- 0:00Introduction to the problem
- 0:23Testing challenges
- 1:20Approaches for test selection
- 4:57Building the prediction model
- 7:13Model implementation
related talks
55mBuilding a Recommendation Engine with Machine Learning Techniques
45mUsing AI, statistics, and set theory to know what to test
18mPredictive Testing in JavaScript with Machine Learning
39mA recommendation engine for your applications
52mTest Impact Analysis. Smarten up your pipeline and reduce t
35mRisk Based Code Coverage Analysis
01h 02mReduce Developer Toil: Autonomous AI for Enterprise Testing @ Scale
45mOptimization of Regression Tests