Creating models for crime prevention

Creating models for crime prevention
For this final assignment, you will have to develop a domain model for crime
(prevention) and design an analysis and support model for it. The assignment consists
of several steps that you have to do. Create a self-contained report – with a brief
introduction, motivation for your choices, captions for graphs, etcetera. Also make sure
that you clearly indicate which question is answered.
For structure, layout and readability you receive 5 points. The maximum length is 1500
words, excluding graphs and tables (see expected length that is specified per question).
Note that a plagiarism check will be applied.
How do crimes occur?
There are many theories that (partly) describe how a criminal action can take place. In
this assignment, we focus on the “Routine Activity Theory” (Cohen & Felson, 1979).
This theory states that for a criminal offense to occur, three elements are important:
 a motivated actor (e.g. somebody with an intention to do a street robbery)
 a suitable target (e.g. somebody with a wallet on the street)
 no social control (e.g., nobody nearby)
The first of these items refers to the background of the criminal, who from a
psychological perspective should be motivated to perform an action that is considered a
crime. A motivated actor has an intention for crime. Such intentions are usually caused
by a combination of desire for valuable things and a belief that criminal action is an
appropriate way of fulfilling this desire. The second and third item in the Routine Activity
Theory refer to the social context. Together they indicate what is called an opportunity
for a crime. In a society with less opportunities, even when motivated criminals are
around, less crime will occur, is the idea. When a criminal offense takes place, it adds
up to the number of crimes in a neighborhood, which contributes negatively to
the safety. Some of the crimes that have happened are reported to the police, leading to
a number of reported cases in a neighborhood.
Question 1 (20 points)
Maximum length, including introduction: graph + 1/2 page
Based on the text above, create a graphical representation of a model that describes
the occurrence of criminal offenses from the perspective of an offender. Use the “belief-desire-intention” (BDI) model as a skeleton. You can use the terms that are printed
in italic as indication for the concepts. Include a box in your model that describes the
boundary between the internal (to the criminal) and external factors. Explain how the
BDI model is visible in your domain model (e.g. by giving the concepts different colors)
and explain the choices and/or assumptions that are relevant for your model.
Question 2 (10 points)
Expected length: 1/4 page
Describe two “characteristic patterns” (in other words: hypothesis about the outcome of
simulations under certain circumstances) that you expect your model to show to be able
to evaluate later whether it is working correctly. Make sure you formulate these
unambiguously and complete.
Question 3 (15 points)
Expected length: two tables
Provide a formalization for the concepts in your model, using predicates and sorts.
Describe it using a table as we have seen in the course. Explain for each of the
concepts why you have chosen this specific formalization. Do this in an extra column in
the table.
Question 4 (10 points)
Expected length: 1/4 page
When using a numerical representation for concepts that are true or false, it is still
possible to use boolean-like calculation rules to specify how the value of another
concept at the next time step is derived from other concepts. Suppose that we have the
following three concepts, which are physiological concepts related to the desire for
aggression in criminals (note: these do not have to be used in the domain model of the
other questions):
 high_level_of_testosteron (0/1)
 low_serotine_level (0/1)
 desire_for_agression (0/1)
Now, create numerical calculation rules (using arithmetic operators like +, -, * and / or
MAX, MIN functions) to define calculation rules to specify the following semi-formal
rules: IF high_level_of_testosteron OR low_serotine_level THEN desire_for_agression
 IF high_level_of_testosteron AND low_serotine_level THEN
desire_for_agression
 IF NOT(high_level_of_testosteron) AND low_serotine_level THEN
desire_for_agression
Hint: you have to create numerical calculation rules of the format: A = B + C, where A, B
and C should be the concepts introduced above and the combination of operators have
to be found by you.
Question 5 (20 points)
Expected length: graph + table + one paragraph
Now assume that we are going to build an intelligent system that will assess whether
the safety in a neighborhood is too low, based on the number of reported cases.
 Create a graphical representation of this model; make sure to include a desire
and assessment.
 Also create a table with the predicates that are used in this analysis model.
 Explain what kind of reasoning is used (forward or backward).
Question 6 (20 points)
Expected length: three figures + one paragraph
The ultimate aim of the intelligent system is to propose actions that reduce the number
of crimes that occur; specifically, to propose actions that prevent crimes from happening
in a neighborhood with a too low safety. According to our theory, there are two possible
actions to prevent crimes: 1) increase the social control by placing camera’s and, 2)
invest in social workers to reduce the belief of potential criminals that a criminal action is
an appropriate way of fulling the desire for valuable things.
 Create an extended domain model by adding these actions to the domain model.
 For both actions, provide a graphical representation of a backward reasoning
support model. You don’t have to create a separate table for the predicates, but
could include the predicates in the graph. Make sure to use a “propose” predicate
and also the right other predicates that are used for backward reasoning support
models.
References[1] Cohen, L.E. and Felson, M. (1979). Social change and crime rate trends: a routine
activity approach. American Sociological Review, vol. 44, pp. 588-608.
Rubric
final assignment
Resit final assignment
Criteria Ratings Pts
This criterion is linked to a learning outcomeQuestion 1: correct graph and
explanation of at least 2 choices. 20 Pts
Full
marks
0 Pts
No
marks
20 pts
This criterion is linked to a learning outcomeQuestion 2: two correct
hypotheses 10 Pts
Full
marks
0 Pts
No
marks
10 pts
This criterion is linked to a learning outcomeQuestion 3: two tables with
correct formalization of concepts and sorts 15 Pts
Full
marks
0 Pts
No
marks
15 pts
This criterion is linked to a learning outcomeQuestion 4: correct formalization
of three calculation rules 10 Pts
Full
marks
0 Pts
No
marks
10 pts
This criterion is linked to a learning outcomeQuestion 5a: correct graph
12 Pts
Full
marks
0 Pts
No
marks
12 pts
This criterion is linked to a learning outcomeQuestion 5b: correct
formalization of concepts 4 Pts
Full
marks
0 Pts
No
marks
4 pts
This criterion is linked to a learning outcomeQuestion 5c: correct explanation
of reasoning method 4 Pts
Full
marks
0 Pts
No
marks
4 ptsResit final assignment
Criteria Ratings Pts
This criterion is linked to a learning outcomeQuestion 6a: correct extended
domain model 5 Pts
Full
marks
0 Pts
No
marks
5 pts
This criterion is linked to a learning outcomeQuestion 6b: two correct support
models with the right predicate names used in the graph 15 Pts
Full
marks
0 Pts
No
marks
15 pts
This criterion is linked to a learning outcomeOverall: well structured, readable
report with appropriate layout 5 Pts
Full
marks
0 Pts
No
marks
5 pts
Total points: 10