Scorecards

Uses applicant/customer data to statistically assess the likelihood of a positive outcome to any financial risk.

Credit scoring uses applicant/customer data to assess statistically the likelihood of a positive outcome to any financial risk.

Credit scoring is based on the premise that behavior of new applicants for credit will closely resemble that of similar past applicants.

Specific scorecards are modeled using the lender’s own data obtained through its own customer base.

Our solutions are based on neural computing. Neural networks are algorithms and data analysis methods. An artificial neural network consists in a large number of simple processing units, linked through weighted connections. The power of neural networks derives from combining multiple units in one single network. An artificial neural network is non-linear, and thus an exceptionally powerful method for real-time data analysis that allows modeling extremely difficult dependencies.

Strengths of neural technology:

  • It has the capacity to learn from experience and apply what has been learnt at the decision moment;
  • A neural system decision is purely objective and not subjective;
  • It can understand relationships between variables hidden deep within data;
  • The ability to cope with little or incomplete data;
  • It develops highly accurate scorecards and models – it reaches more than 90% in accuracy;

One of the most important factors governing accurate scorecard building is the quantity and quality of the input data. The neural approach solves problems in a uniquely different way.  Instead of requiring a human to solve the problem, through a manual statistical analysis and programming processes, the neural computer learns the key relationships underlying historic data.

It generalizes from those relationships and produces predictive models that are used to solve problems. Once the model is built it can be easily updated with new data.