Skip to content
SIH Buddyby Ganeev Singh
Dev
All problem statements
SIH26189Worth consideringacceptance 3/5

AI-Powered Criminal Network Analysis System

Ministry of Home Affairs · Blockchain & Cybersecurity · Software

The link-analysis concept is valuable and demos well, but you synthesise the investigation data so the network only shows what you built in, and commercial tools already exist — make robust entity extraction from messy text your real contribution, since that is the actual bottleneck.

What it actually is

Criminals operate through networks of associates, money, phones and locations, and investigators have the data — FIRs, call records, financial records, surveillance — but it is fragmented and manual analysis misses connections. The ask is a system that extracts entities from these varied sources, builds a relationship graph, identifies key players, and flags suspicious patterns to help investigators see the network.

What to build

A criminal-network analysis system that ingests structured and unstructured law-enforcement data — FIRs and reports, call detail records, financial transactions, surveillance notes — extracts entities like people, phone numbers, vehicles, locations and organisations via NLP, builds a relationship graph linking them across sources, identifies central and influential individuals through graph centrality, detects suspicious patterns and unusual activity, and presents the network visually with analytical insights so an investigator can explore how suspects connect and who the key nodes are.

Smallest thing that wins the room

Ingest a set of mock FIRs and call records, extract the people, numbers and locations, build the relationship graph, and highlight the central individual connecting otherwise separate clusters with the suspicious link pattern that reveals them, all explorable on the graph.

How crowded this one gets

A guess, projected from the 2025 statements — the last year where both the submission counts and the winners were published.

Quiet70–160 teams expectedroughly 1 in 60–139 wins it

Quieter than 78% of the 226 · #50 of 226 by expected field

Few teams are likely to go here. The best odds on the board come from statements like this.

Why: defence, intelligence and space bodies drew small fields.

This is a guess, not a fact

Nobody has published 2026’s numbers yet. This is an analysed estimate from last year’s pattern, so please do not take it as the truth — check the live counter on the SIH portal before you decide anything. The range covers the middle half of likely outcomes, so one statement in two lands outside it. Entry closes at 500 ideas per statement, so no range goes past that — a statement that reaches the cap fills and shuts rather than drawing an unlimited crowd. The model reads only three things a team can see before choosing — software or hardware, the theme, and what kind of body posted it — and those explain about a quarter of the variation in last year’s field sizes (R² 0.25 on held-out statements). Trust the band more than the number, and the ordering more than either. It cannot see how good your idea is, which is the part that actually decides it.

The scores

The number is the shorthand. The line under it is the reason.

What you will be writing

  • NLP entity extraction from reports (spaCy / LLM)
  • Neo4j / NetworkX relationship graph
  • Graph centrality for key-actor identification
  • Call detail record and transaction linking
  • Suspicious-pattern / community detection
  • Link-analysis visualisation (Cytoscape.js)
  • Criminal intelligence
  • Graph analytics
  • Investigative tooling

Prior art to read before you start

entity extraction and relationship mapping · criminal network key-actor detection · multi-source investigative link analysis

Analysed by Claude Opus. Every score above is a judgment call with its reasoning attached — kindly cross-check this against the official statement on the SIH portal before your team commits to it.