Pattern Intelligence is GNI's long-term research hub, designed for analysts and researchers who want to go beyond the latest report and understand how global intelligence evolves over time. Every pipeline run produces structured data — escalation scores, sentiment scores, MAD verdicts, and confidence intervals — all of which are tracked, stored, and available for deep analysis here. The GPVS (GNI Prediction Validation Standard) system continuously scores each agent's directional predictions against reality, building an evidence-based accuracy record over multiple horizons (7d, 30d, 180d). Confidence interval analysis using the t-distribution (t=4.303, n=3, alpha=0.05) provides statistical rigor to every sentiment score, making GNI's outputs IEEE paper-citable. All datasets are available for download via the Export API, enabling full replication of GNI's methodology for academic research and external validation.
0 pipeline runs are stored and available for analysis. Each run captures escalation score, sentiment score, MAD verdict, and confidence interval width. Scroll back through history to identify escalation patterns, trend reversals, and structural shifts over time.
A full research workspace with escalation trend charts, confidence interval width over time, and per-agent MAD accuracy tracking. Compare how Bull, Bear, Black Swan, and Ostrich agents perform across different geopolitical conditions. Designed for deep statistical analysis and IEEE paper evidence gathering.
Cross-horizon Forecast Analysis (CFA) tracks GNI prediction accuracy across three time horizons: 7-day, 30-day, and 180-day. Per-agent accuracy scores reveal which MAD agents perform best under different market conditions. This is the statistical proof layer for GNI's GPVS standard — essential for IEEE paper validation.
Weekly digests compiled every Sunday provide a macro-level view of geopolitical escalation trends. Week-over-week escalation delta is one of GNI's most reliable structural trend signals — consistent increases indicate sustained threat buildup. Use the digest time series to identify multi-week escalation cycles that single pipeline runs might miss.
Documents the complete scientific methodology behind GNI: the confidence interval formula using t-distribution (t=4.303, n=3, alpha=0.05), the Quadratic MAD agent design based on the Johari Window framework, and the GPVS prediction validation standard. All methodology is IEEE paper-citable and designed for academic peer review. Start here if you are writing a research paper that references GNI.