IntelScoring

A system that provides automation of the full model - building cycle, taking its efficiency and speed to a new level. Explore Now
Bringing the solution

"The idea of creating an intelligent automated model building system came to light during forecasting models development for a variety of customers.
The models development itself is an overly labor-intensive and time-consuming process, whether on a client or a consultancy side. This inspired us to bring out a solution that would automate all the routine and bring a vast efficiency improvement to the process overall."

Andrey Matviychuk Founder & CEO at IntelSoft
Key benefits

IntelScoring allows:

 
Higher efficiency of credit risk analysis

Compared to models developed by dedicated experts as it conducts an intelligent search of incommensurably more factor combinations and variables categorization methods

Significantly improved model building speed

Making models more adaptable to changing environments

Considerable cost cuts

In maintaining a large number of internal staff or consulting services



 
System assignment
System highlights

IntelScoring combines:


  • Classic tools

    • Linear and logistic regression
    • Classification trees
    • Discriminant models
    • Other
  • Artificial intelligence methods

    • Fuzzy logic
    • Neural networks
    • Genetic algorithms
  • Efficiency criteria

    • Gini coefficient
    • Kolmogorov-Smirnov statistic
    • Type I and type II errors under a variety of separator levels
    • Other
Additional features

Along with full model building automation, IntelScoring enables analysts to intervene in the process at any stage (according to the bank’s corporate regulations and access level).
For example, a designated employee can build a number of models for different customer groups and loan products along with being able to:

  • Define the sampling time frame for tuning and testing a model;
  • Manually adjust (if needed) category ranges following an automated binning;
  • Compulsorily select any factors considered to be the most important for the risk assessment process.

Technical aspect


 
The system is based on a client-server architecture

At this stage, Web browser serves as a client enabling to use the application on both PC and mobile devices (with future possibility of iOS and Android apps support)

The server-side is powered by .Net Core 2.0 platform enabling a cross-platform execution on both Windows and Linux servers

System deployment:
  • - On-premice
  • - Cloud (Azure, Amazon etc.)

Contact Us

For more information or a live demo, please contact us any time.

Sending...
Something went wrong. Please try again.
Your message was sent, thank you!
or Email Us At
hello@intelsofttech.com