IMSI-TRACK ML
From Detections to Intelligence.
IMSI-TRACK ML adds a machine-learning intelligence layer on top of IMSI-TRACK PRO. It correlates identities, learns normal behavior, flags anomalies and predicts presence — turning millions of raw detections into ranked, evidence-linked leads, with correlation latency under 2 seconds.
The Analyst That Never Sleeps
Detection tells you a device is present; intelligence tells you which ones matter. IMSI-TRACK ML profiles every identity's pattern of life, links devices that move together, and surfaces the anomalies — new SIMs, burner swaps, coordinated presence — before an analyst would ever scroll to them.
Behavioral Profiling
Builds a per-identity pattern-of-life (times, zones, dwell, co-presence) and scores deviation continuously, flagging first-seen devices and unusual movement within 2 s of the triggering detection.
Device Correlation
Links IMSI/IMEI pairs and devices that repeatedly co-locate, exposing SIM swaps, multi-handset users and networks of contacts with a confidence score per link.
Predictive Presence
Forecasts when and where a watched identity is likely to appear over a configurable 15–120 min horizon, prioritizing sensor attention and alerting operators ahead of events.
Watchlists & Whitelisting
Auto-classifies staff/known devices to suppress noise and escalates watchlist hits, cutting false alerts and focusing operators on ranked, evidence-linked leads.
The Result
- Correlation and anomaly scoring < 2 s per detection
- Per-identity pattern-of-life with confidence-scored device links
- Predictive presence over a 15–120 min horizon
- Watchlist escalation + auto-whitelist to cut false alerts
Technical Specifications
Analytics Engine
| Correlation latency | < 2 s per detection |
| Models | Behavioral profiling · anomaly detection · link analysis · forecasting |
| Anomaly triggers | First-seen · SIM swap · unusual zone/time · co-presence |
| Confidence scoring | 0–100% per link and per alert |
| Prediction horizon | 15–120 min (configurable) |
Ordering Information
Part numbers and configuration options
| Part Number | Description | Availability |
|---|---|---|
| SI-ITM-ENG | ML intelligence appliance · 1U | Standard |
| SI-ITM-SW-ENG | ML engine license (PU co-resident) | Standard |
| SI-ITM-SW-PRED | Predictive-presence module | Optional |
| SI-ITM-SW-WL | Watchlist & classification module | Optional |
| SI-ITM-SUP-24 | 24-month support & maintenance | Optional |
Implementation Process
Site Survey
- Confirm IMSI-TRACK PRO baseline
- Define zones of interest
- Agree watchlist and privacy policy
System Design
- Appliance vs co-resident decision
- Model configuration
- Alert-ranking policy
- Integration points
Installation
- Appliance/license commissioning
- Connection to PRO data feeds
- Dashboard setup
Calibration
- Model training on historical data
- Anomaly threshold tuning
- Whitelist seeding
Training
- Analyst workflows
- Dossier and reporting
- Watchlist management
- Alert triage
Go-Live
- Live intelligence enabled
- Ranked alerts to operators
- Support transition
Typical deployment timeline: 4-8 weeks depending on facility size and complexity
Frequently Asked Questions
Put an Analyst on Every Detection
Request a demonstration of correlation, anomaly scoring and predictive presence.
IMSI-TRACK PRO
the detection and geolocation base layer this intelligence engine runs on.