The Identification of Sequential Patterns Preceding the Occurrence of Political Events of Interest

In this paper we present an approach to identifying patterns of behavior preceding political instability events in countries from sampled factor data. This process is based on the concept of state-space back-chaining. A list of sampled, quantized factor data sampled over a range of discrete times define a ”state-space” of a country, and the list of quantized factor data associated with a country at a particular instance in time defines a state. The state of a country changes over time and instances of two countries changing in the same manner define a pattern. At some country states political instability events of interest (such as the onset of regime change, insurgency, ethnic violence, etc) may be observed to occur. We discuss a method to identify the set of factors which define a state space and pattern for combinations of selected political instability events of interest such as rebellion, insurgency, civil war, etc. This approach can be used to identify the observable behaviors before the occurrence of events of interest. The backwards chaining methodology is implemented in the Java programming language and run over a set of factor data. We present a pattern for domestic political crisis found using this method.