JCAT scenario and planning outcome analysis software offers a powerful and user-friendly, software for uncertainty and what-if analysis of complex plans and scenarios.
Bayesian techniques can provide powerful analytical and predictive capabilities; however, true Bayesian analysis, with its voracious appetite for numbers, can quickly overcome analysts and computers attempting to process complex situations. When a temporal element is added, the complexity increases even further, rendering most Bayesian analysis techniques ineffective either through overwhelming mathematical requirements or the necessity to grossly oversimplify the model.
The US Air Force Research Laboratory Information Directorate (AFRL/RI) has developed and continues to improve JCAT (Java Causal Analysis Toolkit). JCAT allows analysts to provide exact model parameters for important elements while estimating the remaining factors using causal (not statistical) reasoning. The mathematical basis behind the causal estimations, “Recursive Noisy OR,” maintains a theoretically sound implementation of Bayesian theory while also simplifying the process by incorporating a semantic approach to model development. The end result is a user-friendly software tool with the power to tackle complex analytical tasks modeled dynamically through time.
Notable implementations of JCAT include a detailed analysis that predicted unrest in Thailand in 2010, and a complex model to identify potential outcomes of missile defense politics and policy. Use of JCAT within the US government is growing, and the core elements of the software are potentially readily adaptable to a broad range of analytical applications and techniques.
BENEFITS:
- Powerful: Able to process large models with temporal connections spanning significant time lengths where competing techniques cannot
- User-Friendly: JCAT simplifies the modeling process, making meaningful analysis possible for users lacking an intense mathematical background; JCAT makes it feasible and relatively easy for any analyst to develop and assess complex models
- Broadly Applicable: Wide potential for use, including marketing/communication campaign and financial market analysis, to investigations into deconflicting airspace that would improve civilian air travel