Lecture 23 Simulation Crime and Crime Pattern Using Cellular Automata and GIS - PowerPoint PPT Presentation

Loading...

PPT – Lecture 23 Simulation Crime and Crime Pattern Using Cellular Automata and GIS PowerPoint presentation | free to download - id: 4c0f7-NTQxN



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Lecture 23 Simulation Crime and Crime Pattern Using Cellular Automata and GIS

Description:

23-4 Tension ... Tension serves as the CA state variable. ... will increase, as tension increases, in order to prevent the reoccurrence of crime. ... – PowerPoint PPT presentation

Number of Views:319
Avg rating:3.0/5.0
Slides: 20
Provided by: junl4
Learn more at: http://www.unc.edu
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Lecture 23 Simulation Crime and Crime Pattern Using Cellular Automata and GIS


1
Lecture 23 Simulation Crime and Crime Pattern
Using Cellular Automata and GIS
  • 23-1 Spatial criminology
  • Early studies were limited to macro level
    approaches. Crime facts were aggregated at the
    neighborhood level. In the 19th century, both
    offender and target (or place) were aggregated at
    different levels - neighborhood, cities,
    counties, or states. At this period, little was
    done to reveal the relationship between specific
    places and crime.
  • Since the 1960s, criminologists have developed
    theories to understand crime events at the micro
    level a place, a target, or even a street
    address.

2
23-1 Spatial criminology (Cont.)
  • From the view of environmental criminology,
    criminal events must be understood as results of
    offenders, victims or criminal targets, and laws
    in specific setting at particular times and
    places.
  • RAT studies crime events at the micro level,
    which is even more detailed than environmental
    criminology. Four conditions must be met for a
    successful crime event a motivated offender a
    desirable target occupying the same place and
    time and controllers (such as handlers,
    guardians and place managers) must be absent.

3
23-2 Crime Patterns
  • Basic facts about crime patterns
  • Crime events are rare. The chances of a crime at
    a specific location during a given short time
    interval is very small.
  • At any geographic scale, many units have few or
    no crime events and a few have many crime events.
  • At any temporal scale, many time intervals have
    few crime events and a few have many crime
    events.

4
23-2 Crime Patterns (Cont.)
  • Retrospectively, crime patterns are easy to
    detect.
  • The larger the area selected or the longer the
    time interval selected, the more obvious the
    crime pattern.
  • Prospectively, precise crime patterns are
    difficult to anticipate but the vaguer the
    pattern that is acceptable, the easier it will be
    predict.

5
23-3 A Probabilistic Approach of RAT
Situation S is defined as (t time, i place, j
offender, k offense, v person). Targets
T Desirability ? Guardians G Capability
? Offenders O Motivation µ Handlers
H Intimacy ß Places P Accessibility
? Managers M Effectiveness ?
6
23-3 A Probabilistic Approach of RAT (Cont.)
  • The possibility equation can be used to evaluate
    crime pattern at time t. Limitations
  • Static model
  • Local model
  • Could not simulate interactions between offenders
    and targets
  • Could not reveal relationships between crime
    patterns

7
23-4 Tension
To figure out the transition function, we need to
understand the changes of interaction between
offenders and targets over space and time, and
also interaction among targets, and offenders.
8
23-4 Tension (Cont.)
  • In order to let targets interact with
    offenders/other targets over space, the concept
    of tension has been developed in this
    research. Generally, tension is crime
    anticipation of targets.

9
23-5 Modification of the crime likelihood formula
Since only commercial property robbery is
considered here, we may simplify the original
formula Offense (k) and person (v) have been
removed from likelihood formula. Also victims,
guardians, and managers are incorporated into the
place, and there is no reason to track individual
people. Place accessibility (?) has also been
dropped for simplicity, and because it could be
incorporated in the concept of place management.
10
23-5 Modification of the crime likelihood formula
(Cont.)
  • Assumptions about offender movement and target
    reactions to crimes -
  • Offender movement is assumed to follow Monte
    Carlo simulation, and
  • Targets adjust their protection level against
    offenders according to their expectations of
    crime. Target reaction follows the CA model.

11
23-5 Modification of the crime likelihood formula
(Cont.)
  • Thus, the offender factor has been isolated from
    (3) and simulated separately. Without the
    offender factor in (3), the likelihood of crime
    is changed to protection index (PVI) of target.

12
23-5 Modification of the crime likelihood formula
(Cont.)
  • A new crime likelihood formula (5) - Ln(S)ct is
    the modified crime likelihood evaluation
    equation. The modified crime likelihood will
    compare the value of the motivation and
    prevention index. If is greater than PVI, then a
    crime will occur.

13
23-6 Offender Movement Simulation
  • The movements of offenders are simulated with a
    Monte Carlo simulation. The probability of an
    offender appearing at a particular target is
    inverse to its distance from the target. The
    probability matrix of contacts (presence of
    offenders at targets) between target group and
    offender group forms the MIF (Mean Information
    Field). MIF is used to determine offenders
    movements. The shorter the distance between a
    target and an offender the higher the probability
    of contact.

14
23-6 Offender Movement Simulation (Cont.)
  • Two conditions are required for crime to occur at
    a target
  • Presence of at least one offender
  • Ln(S)ct is positive, otherwise the crime will not
    occur.
  • Whether a crime occurs successfully or not is
    determined by another random barrier () and
    Ln(S)ct.

15
23-7 Target Reaction CA Model
  • State Variable
  • Tension serves as the CA state variable. Other
    target variables, which include desirability,
    capability, and effectiveness, can be calculated
    from tension directly or indirectly, based on
    their relationship with tension.

16
23-7 Target Reaction CA Model (Cont.)
  • The transition function f varies under three
    different situations
  • No crime occurs successfully and significant
    tension difference exists between cell c and its
    neighbors
  • No crime occurs successfully and no significant
    tension difference exists between cell c and its
    neighbors
  • Crime occurs successfully.

17
23-7 Target Reaction CA Model (Cont.)
  • The relationship between TS and d is assumed to
    be linear

18
23-7 Target Reaction CA Model (Cont.)
  • Management is assumed as a positive function of
    target desirability. Greater property value
    indicates greater desirability, and greater
    return on investment, indicating higher
    management effectiveness.

19
23-7 Target Reaction CA Model (Cont.)
  • In crime theory, guardianship is defined as a
    function of victimization of the place. ? will
    increase, as tension increases, in order to
    prevent the reoccurrence of crime.
About PowerShow.com