Skip to main content

Risk management


At your request, we will design a solution tailored to your needs.


Discover the potential of this application from Clementine staff!


Clementine experts can even provide on-site training to help your make the most of your applications.


Clementine events present the latest trends and developments in data science and natural language processing.

Risk management


Risk management in financial institutions

Banks must comply with increasingly strict risk management rules, and the use of complex mathematical and statistical methods is essential for a significant part of the minimum capital requirement calculations of the Basel II-IV system.

Our experts supported the development and implementation of risk management models at several different financial institutions. CLEMRISK is the development of our experts with many years of experience and business analysis experience, which is based on the most modern data analysis tool, IBM SPSS Modeler, which is used in many Hungarian banks for data analysis and modeling.

CLEMRISK can create two main types of scorecards

Application score: when applying for a new loan, we assess the risks inherent in the client and the transaction, rate the client, and then decide on the acceptance of the application and its conditions.

Behavioral score: during the term, we continuously monitor the customer's behavioral habits (e.g. late payment) and determine the risks of the transactions. The expected loss can be calculated from the risks, which is a determining factor during the creation of provisions.

ClemRisk guides you through the entire process of risk management modeling:

  • Data quality assessment
  • Handling of missing and outlier data
  • Data transformations
  • Creating a test environment
  • Choosing a modeling technique, building models
  • Comparison of models, evaluation


CLEMRISK uses the modeling algorithms found in the wide model palette of IBM SPSS Modeler, which includes many models, from regression-type models widely used in the banking environment, to decision trees, to neural networks.


Perhaps the most specific area of risk management modeling is model evaluation. To evaluate the models, we use performance indicators that not only show the differentiating ability of individual scorecards, but also enable comparisons between different scorecards.


It is important to integrate and automate the prepared solution into everyday business processes, from data transformations to automated reporting and customer classification.

Legal requirements

Since the range of risk management-type data is particularly sensitive to data protection principles, as well as the operation of financial institutions, their risk management systems are regulated by strict regulations, therefore we take into account the relevant current legislation, government decrees and MNB recommendations in every case and in every step of the risk management analysis.