Predictive Analytics Engine (PAE)

Network operational risks are those associated with virus attacks, targeted attacks (hacking) and physical attacks (damaging or immobilizing technology infrastructure).

Predictive Analytics Engine (PAE) uses quantitative modelling techniques, enabling a quantification of risk metrics for such attacks. These are then utilized in the calculation of the cyber threat Value at Risk, risk-capital measures and also the associated cost of mitigation insurance.

For the first time, organizations have the financial loss exposure, caused by cyber threats actually experienced, available to risk manage such potential financial losses. Predictive analytics of each organizations’ network attack data creates forward looking financial values at risk, facilitating pro-active cyber risk management strategies and pre-emptive actions to be formulated.

Having current and future predictive values provides the means to evaluate capital allocation efficiencies for cyber threat management. Having a current I.T. security capability today does not mean it will remain static against future cyber threats. PAE enables forecasting of security requirements into future periods.

PAE analytics provide greater stochastic modelling capabilities than those within n-ORM and are able to compute a wider range of analytical measures aimed at meeting new and emerging requirements for stress testing of risk models.

The system comprises a primary three-phase approach to modelling, with these being a time-series component, a risk calculation component and a post processing layer. Within phase one, there are a number of optional features that may be enabled or disabled by each training programme participant, these being:

• Utilizing a linear or an exponential process model
• A normal or weighted data model whereby the most recent data has a higher degree of importance in the forecast and simulation
• A standard least squares or a robust model to take account of the particularities of cyber attack data.
• Within the second phase, a Monte Carlo simulation model is utilized which takes a range of input and configuration data and computes risk distributions.

The calculation engine generates probability distribution functions, enabling various statistics to be drawn and utilized within the system in deriving the financial quantification of the cyber threats experienced by individual organizations as those for future periods.