Prediction of Telecom Customer Churning

Benchmark vs. Logistic Regression, Random Forest, Boosted Tree & Neural Network

Design an AI-based decision-making system that accurately evaluates customer attrition, to assign a churning score to a customer in a rational way and adapt anti-churning strategies.
Goals & benefits

Identify parameters actually involved in customers’ decision to unsubscribe and help analysts understand cause-and-effect relationships between specific characteristics, their combination, and a churning risk.

Help the CRM team focus only on meaningful cases and take earlier and more personalized anti-churn actions thanks to rapid, systematic, and explainable alerts.

Reduce the turnover of Telecom company’s customers.

  • The top-model is a decision system composed of 33 gradual rules without chaining.
  • Each rule uses from 1 to 7 predictors among the 10 variables that XTRACTIS automatically identified as significant (out of the18 features characterizing consumers).
  • Only a few rules are triggered at a time to compute the decision.

It has a quite good Real Performance (on unknown data).

It computes real-time predictions up to 70,000 decisions/second, offline or online (API).

UC22 scores graph

LoR=Logistic Regression
RFo=Random Forests
BT=Boosted Trees
NN=Neural Networks

Detailed results and explanations in full document

Use Case 2024/07 (v1.1)

Powered by XTRACTIS® REVEAL 13.0.456104 (2023/07)


  1. Problem Definition
  2. XTRACTIS-induced Decision System
  3. XTRACTIS Process
  4. Top-Model Induction
  5. Explained Predictions for 2 unkown cases
  6. Top-Models Benchmark
  7. Appendices