XTRACTIS for Human Resources
Discovery of Discriminatory Biases in the Professional Evaluation of Employees
Benchmark vs. Logistic Regression, Random Forests, Boosted Trees & Neural Networks
This use case is an illustrative example of XTRACTIS’ ability to reveal conscious or unconscious biases in high-risk critical decisional processes.
Goals & Benefits
Identify the specific parameters characterizing each employee and which are significant in his or her manager’s evaluation process.
Improve Human Resources management knowledge by helping companies understand the cause-and-effect relationships between these parameters and managers’ decision strategies.
Give the regulator a tool to check social compliance on a case-by-case basis.
XTRACTIS-induced Decision System
Intelligible Model, Explainable Decisions
- The top-model is a very simple decision system composed of 6 binary rules without chaining.
- Each rule uses from 1 to 3 predictors among the 4 variables that XTRACTIS automatically identified as significant in the decision process (out of the 20 Potential Predictors).
- Two rules are discriminatory as they clearly state that a a specific origin of the first name is a systematic decision criterion.
High Predictive Capacity
It has a perfect Real Performance (on unknown data).
Ready to Deploy
It computes real-time predictions up to 70,000 decisions/second, offline or online (API).
Intelligibility x Performance Scores Benchmarks
LoR = Logistic Regression RFo=Random Forests
BT=Boosted Trees
NN=Neural Networks
Detailed results and explanations in full document
Use Case 2024/11 (v1.1)
Powered by XTRACTIS® REVEAL v13.0.50246 (2024/04)
CONTENTS
- Problem Definition
- XTRACTIS-induced Decision System
- XTRACTIS Process
- Top-Model Induction
- Explained Predictions for 3 unkown cases
- Top-Models Benchmark
- Quantitative Metrics
