Dynamic Forecasting Conditional Probability of Bombing Attacks Based on Time-Series and Intervention Analysis.

Area of Science:

• * Criminology and Political Science
• * Data Science and Time-Series Analysis
• * Global Security Studies

Background:

• * Bombing attacks are the most frequent terrorist tactic globally, according to the Global Terrorism Database (GTD).
• * Understanding and forecasting bombing attack probability is crucial for security and policy.
• * Previous analyses have not adequately addressed the dynamic forecasting of bombing attack probabilities.

Purpose of the Study:

• * To develop and validate a time-series model for analyzing and forecasting the conditional probability of bombing attacks (CPBA).
• * To identify and analyze sudden increases in CPBA using intervention analysis.
• * To forecast CPBA trends for future security planning.

Main Methods:

• * Development of an integrated time-series and intervention analysis model.
• * Application of the model to historical data from the Global Terrorism Database (GTD) (2004-2014).
• * Validation of the model using 2014 data, achieving an average relative error of 3.5%.

Main Results:

• * A significant, dramatic increase in CPBA was observed in late 2011, peaking at 16.0% over two months.
• * Post-intervention, CPBA remained 9.0% higher than predicted levels after temporary effects decayed.
• * No significant fluctuations were detected in the conditional probability of armed assaults.

Conclusions:

• * Social unrest, such as troop withdrawals from Afghanistan and Iraq, may correlate with increased CPBA in the region.
• * The developed model provides a reliable method for forecasting CPBA up to 2064.
• * The findings offer valuable insights for counter-terrorism strategies and resource allocation.