Predictive Policing System Using ETAS Point Process Model

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current crime forecasting methods, such as hotspot policing and kernel density mapping, fail to accurately predict future crime occurrences by ignoring spontaneous and triggered events and lack a rigorous methodology for parameter estimation, leading to inefficient allocation of police resources and high variance in risk estimates.

Innovation Solution

A predictive policing system utilizing an ETAS point process model that processes historical crime data to assign probabilities to geographic regions, providing real-time crime forecasts and visual maps for effective resource allocation, incorporating GPS feedback for officer safety and using a cloud-based platform for scalability and user-friendliness.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If crime hotspot maps are used to quantify future crime risk, then police resource allocation is improved, but the accuracy of prediction is reduced because background events are ignored

Engineering Contradiction:

Improvepolice resource allocation efficiencyVSAvoidcrime prediction accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent segments crime events into two distinct categories: background events (spontaneous crimes not triggered by prior events) and triggered events (crimes resulting from prior crimes). This segmentation allows the model to separately estimate and combine the risks from both sources, thereby improving prediction accuracy while maintaining efficient resource allocation through hotspot mapping.

Inventive Principle:

Principle #1Segmentation

2Ease of operation

If crime hotspot maps assume short term trends persist into the future, then ease of operation is improved, but prediction reliability deteriorates due to ignoring background events

Engineering Contradiction:

Improvemodel simplicityVSAvoidprediction reliability

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The patent implements a dynamic model where the crime risk is not static but evolves over time through the interaction of background events and triggered events. The model dynamically updates risk estimates by combining the persistent short-term trends (from triggered events) with the stochastic background event rate, thereby maintaining operational simplicity while improving prediction reliability.

Inventive Principle:

Principle #15Dynamics

3Measurement precision

If sophisticated computer models are used to assign probabilities to space-time regions, then crime prediction accuracy is improved, but device complexity increases

Engineering Contradiction:

Improvecrime prediction accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces complex mechanical or computational systems with a statistically-based point process model. By using probabilistic mathematics to model crime as a combination of background events and triggered events, the system achieves high prediction accuracy without requiring sophisticated computer models or complex algorithms, thereby reducing system complexity.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11436510B1Event forecasting system

Publication Date: 2022.09.06 PREDPOL INC

AI summary

Generally provided herein is a predictive policing system including at least one crime prediction server constructed to process historical crime data to produce a crime forecast assigning at least one geographic region to at least one crime type for use in crime prevention, deterrence, and disruption practices.