Neural Risk Management

Compania Neural Risk Management a fost infiintata in 2002 ca raspuns la nevoia tot mai mare a pietei financiar-bancare, in curs de dezvoltare, de solutii noi si performante de modele de risk management.

Technology

Our solutions utilize Neural Technologies' proprietary AMAN™ engine.

AMAN™ is a self-growing meta-neural network that combines many types of neural architectures. It can cope with large amounts of data, both numeric and symbolic, and can be implemented from a desktop PC, to a client/server, mainframe or over a distribution environment such as the Internet. This is the core product that can be applied to different industries and functionalities.

What are neural networks?

A neural network is a computing paradigm that is loosely modeled after cortical structures of the brain.

Tasks to which artificial neural networks are applied for:

  • Forecasting financial institutions customers’ solvability;
  • detecting card fraud (used by VISA), etc.;
  • evaluating insurance policies prices;
  • weather forecast;
  • detecting natural resources based on ground characteristics;
  • forecasting value of obligations;
  • forecasting the moments for buy/sell of stocks.

The original inspiration for the technique was from examination of the central nervous system and the neurons (and their axons, dendrites and synapses) which constitute one of its most significant information processing elements. The main feature of such systems is the possibility of learning using examples and experience for improved performance.

Structure

Though is has the same principles of functioning as the human brain, neural networks use a different structure. A neural network is simpler than human brain, but same as the human brain, uses powerful calculation units.

Neural networks characteristics:

Artificial neural networks can be characterized by three elements:

  • model adopted for processing,
  • particular structure of interconnections (architecture)
  • mechanism of adjusting connections (learning paradigm)

Learning algorithm

The main difference between neural networks and other artificial intelligence based systems is the learning paradigm by interacting with environment and increasing performance.

A correct presentation of the informations, witch allows interpretation, forecasting and the response at an external stimulus, allows the network to build a model of the analyzed process. This model can respond can respond to unutilized stimuli in main training process. The information utilized in learning process can be: available informations or in-out couple (witch use relations like cause-effect).

It is known worldwide that the most performant scoring models and risk management solutions are developed using both artificial intelligence and the professionalism of human expert.

Scorecards developed by Neural Risk Management use both the power and accuracy of neural networks and the expertise of domain professionals, which is why our product quality meets the highest domain standards.

Neural computing: strengths

  • Fast and flexible
  • Needs less developing time than conventional systems
  • The best instrument for identifying models and trends from a great variety of data type and volume
  • High quality

Products developed using neural technology and Neural Technologies' experience are characterized by high data administration and high precision results.

A neural network is a computing paradigm that is loosely modeled after cortical structures of the brain.