Methodology

Event Classification Framework

Articles are ingested from a curated set of sources. A structured classifier assigns an event type, severity (1–5), and a confidence score (0.0–1.0). The intent is consistency and auditability rather than prediction.

Severity scoring

• 1 — Routine developments with limited operational impact.
• 2 — Notable developments; localized or low-intensity effects.
• 3 — Material escalation, sustained disruption, or major policy action.
• 4 — Significant escalation with broad operational or strategic consequences.
• 5 — High-intensity conflict or systemic disruption with strategic impact.

Confidence scoring

Confidence reflects the system’s certainty in the structured labels, based on clarity and corroboration in the source text. Low confidence does not imply the event is unimportant; it indicates ambiguity or limited detail.

Event type taxonomy

Event types are mapped into stable buckets (e.g., kinetic, diplomacy, sanctions, procurement, cyber). Taxonomy can expand over time without changing historical records.

Risk Index Construction

Risk indices are aggregated from structured events over rolling windows. A baseline approach is:

• Severity × confidence aggregation to reduce over-weighting uncertain classifications.
• Regional buckets (e.g., global, NATO, BRICS, Indo-Pacific) derived from deterministic country/regional groupings.
• Rolling windows (daily and weekly) to support consistent comparisons over time.

Market Linkage Model

Market analysis is descriptive and window-based:

• Weekly return is computed from the first and last close inside a weekly window.
• Correlation uses Pearson correlation between risk index series and market return series where sufficient data exists.
• Limitations: correlations are not causal and may be unstable across time periods and regimes.

Weapons-to-Company Mapping

Weapons systems referenced in events can be linked to manufacturers and (where applicable) representative tickers. The mapping is curated and scope-limited; it does not imply procurement, revenue share, or exposure magnitude.

Limitations & Disclaimers

• Not financial advice; outputs are informational and descriptive.
• Data coverage varies by source availability and publication practices.
• Automated classification can be uncertain; confidence scores are provided for transparency.