XTRACTIS FOR PREDICTIVE FINANCIAL RISK ANALYSIS
Prediction of One-year Bankruptcy of Companies
Benchmark vs. Logistic Regression, Random Forests, Boosted Trees & Neural Networks
How to predict the bankruptcy of a company in one year automatically, transparently, and successfully from its annual balance sheet?
Help bankers measure the risks incurred when granting loans or payment terms to their business customers.
Find the really influential parameters to evaluate the financial situation of a company.
Guide companies in improving their financial profile.
We obtain a Predictive Model that is:
Intelligible.
A Decision System composed of 14 unchained gradual rules, each using some of the 15 variables that XTRACTIS identified as significant (out of 64 potential predictors characterizing each company financial situation).
Quite Robust.
Comparatively, fairly good real performance in External Test.
Efficient & Operational.
Running in real-time up to 70,000 predictions per second (i7 @2.5GHz with 8 physical cores), offline or online (API).
Use Case 2022/05 (v1.4)
Results by
XTRACTIS® GENERATE 11.3.40062 (2021/11)
DOCUMENT CONTENTS
- Problem Definition
- Xtractis Solution
- Top-Model Induction
- Explained Predictions for 3 cases
- Top-Models Benchmark