... Trough ingenious Customer Intelligence to superb performance ... |
Methodology
The PRO-CR modeling process is based on multiple analytical methods integrated into a single standardized multi level modeling solution.
The analytical approach is a combination of several methods of machine intelligence. Appropriate methods are found in a fully automated optimization process. The principles are pattern cognition and self learning, similar to the way human beings and animals think and work. The solution detects similarities and deviations in an autonomous process and quantifies related risk and opportunities.
The strength of the methodology is that it takes into consideration the whole complexity of a task and there is no reduction as in traditional statistic oriented approaches. The method detects not only exact matches but also cases that are similar to known deviations or characteristics. For example in the credit rating process PRO-CR detects similar “Bad” cases to those which already have been identified as “Bad” and optimizes the early detection of future potential defaults.
The method is tuned to analyze large volumes of data and effectively detect the relevant cases. Thus the management can focus on the essentials. The other important aspect of the methodology is the ongoing learning process. There is an ongoing calibration process to increase the detection quality based on the feedback of detected cases.