SlideShare a Scribd company logo
1 of 23
Analysing a Complex Agent-Based Model 
Using Data-Mining Techniques 
Claire Little, Bruce Edmonds 
Centre for Policy Modelling 
Manchester Metropolitan University 
Ed Fieldhouse, Laurence Lessard-Phillips 
Institute for Social Change 
University of Manchester
An “Inconvenient Truth” 
• That the universe is not arranged for our benefit 
(as researchers studying it) 
• in other words, that assumptions such as the following 
are likely to be wrong: 
– Our planet is the centre of the universe 
– Risky events follow a normal distribution 
– Humans act as if they followed a simple algorithm 
– Society can be understood using simple, universal prinicples 
– etc. etc. 
• In particular, the assumption that there will always (if 
you look hard enough) be models that are: 
– (a) simple enough for us to understand and 
– (b) adequate to what we want to model 
• …is wrong 
Complexity and Context-Dependency, Bruce Edmonds, ECCS, Lisbon, Sept 2010. slide-2
The Alternative 
• Thus consider the alternative, more realistic, 
situation where one is facing some phenomena 
where any model that is adequate (w.r.t. our goals) 
will be too complex for us to completely understand 
• Instead of indulging in wishful thinking this paper 
looks at ways forward under complexity 
• In other words, if we have a simulation model that is 
too complex to completely understand, how can we 
obtain some useful understanding of its 
properties… 
• …and hence use it to leverage some 
understanding/control over the target phenomena 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 3
The Broad Idea 
1. Make relatively complex simulations based 
on available evidence (a “KIDS” approach) 
2. Analyse this simulation in a number of 
ways, including… 
3. Data mining output data in a more holistic 
manner over a broad “space” of settings 
4. Look for patterns in the data that suggest 
(maybe context-dependent) hypotheses 
5. These are candidates for simpler (but 
maybe partial) models of the simulation 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 4
The Model 
• To explore the complex 
mix of factors, 
structures and 
processes that affect 
whether people vote 
• An agent-based model, 
with demographics and 
dynamic social 
networks 
• Was formulated using a 
mixture of qualitative, 
survey data and others’ 
expert opinion/results 
Class 
Age 
Activities 
Ethnicity 
Etc. 
Level-of-Political-Interest 
A Household 
Discuss-politics-with person-23 blue expert=false 
neighbour-network year=10 month=3 
Lots-family-discussions year=10 month=2 
Etc. 
Memory 
An Agent’s Memory of Events 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 5
Overall Structure of Model 
Underlying data about 
population composition 
Demographics of people in 
households 
Social network formation and 
maintenance (homophily) 
Influence via social networks 
• Political discussions 
Voting Behaviour 
Input 
Output
Technique 
• Instead of initiating ‘thin’ analyses of the 
simulation behaviour (e.g. 1 or 2D parameter 
sweeps/correlation models against a few key 
output measures) 
• To sample a multi-dimensional space of 
settings and cluster on a multi-dimensional 
space of output indicators (in this case 9 
parameter x 13 output measures) 
• Look at the patterns between clusters for 
indications as to hypotheses of behaviour 
• Then test these with targeted simulation 
experiments 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 7
More Holistic but Less Detailed 
1. Many Runs, Randomly 
Sampling Parameter Values 
Multi-Dimensional 
Space of Parameter 
Settings 
2. Analyse data set of result 
measures using data mining 
3. Look for patterns that you 
might then check in a more 
systematic manner 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 8
Parameter ranges 
3862 independent runs with parameters sampled 
from the following, uniform distributions: 
• density: [0.65, 0.95] 
• drop-activity-prob: [0.05, 0.15] 
• drop-friend-prob: [0, 0.01] 
• emmigration-rate: [0 ,0.03] 
• immigration-rate: [0, 0.02] 
• int-immigration-rate: [0, 0.02] 
• majority-prop: [0.55, 1] 
• prob-move-near: [0, 1] 
• prob-partner: [0.01, 0.03] 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 9
For each of these runs… 
• Measure many different indicators of the outputs 
(say at the end of the simulation) including: 
– Pop.size – population size 
– Av.age – average age 
– Av.adfriends – average number of friends (adults only) 
– Prop.maj – proportion of the majority population 
– Prop.adult – proportion that is adult 
– Prop.1stgen – proportion that are 1st generation immigrant 
– av.clust – average proportion of friends who are friends 
– av.sim.hh – average similarity within households 
– av.sim.fr – average similarity between friends 
– ncvs.ac – number of conversations over activity links 
– ncvs.sc – number of conversations over “school” links 
– Prop. Adults with highest level of political interest 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 10
Dendrogram of hierarchical 
clustering of simulations 
	 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 11
A heatmap of the hierarchical 
clustering 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 12
The within group sum of squares against the 
number of clusters for 10 randomly initialised 
runs using k-means 
	 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 13
Clustergram of PCA-weighted mean of k-mean 
clusters vs. number of clusters 
	 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 14
Centroid plot against the first two discriminant 
functions showing the 3 clusters 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 15
Details of the centroids of the 3 k-means 
clusters 
Attribute Cluster 1 
(543 
records) 
Cluster 2 
(1333 
records) 
Cluster 3 
( 1986 
records) 
Pop.size 100 557 1750 
Av.age 76 58 55 
Av.adfriends 0.73 1.36 1.82 
Prop.maj 74% 67% 65% 
Prop.adult 99% 94% 93.5% 
Prop.1stgen 8% 13% 14% 
av.clust 0.97 0.84 0.70 
av.sim.hh 2.45 3.53 3.74 
av.sim.fr 2.82 3.70 3.33 
Rate ncvs.ac 1.3% 1.3% 0.0% 
Rate ncvs.sc 0.45% 0.20% 0.13% 
Prop. Adults 
0.97% 1.6% 1.7% 
involved 
Within cluster 
sum of squares 
6748.243 11288.460 7407.591 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 16
the 3 clusters against the parameters: emigration 
rate, immigration rate, internal immigration rate 
	 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 17
Pop Size Av. Age 
	 
Multi-Dimensional Scatter Graphs 
Av Sim Hh Av Sim Fr 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 18
Average proportion of similar friends 
against time for different immigration 
rates 
0.95	 
0.9	 
0.85	 
0.8	 
0.75	 
0.7	 
0	 6	 12	18	24	30	36	42	48	54	60	66	72	78	84	90	96	 
0	 
0.0025	 
0.005	 
0.0075	 
0.01	 
0.0125	 
0.015	 
0.0175	 
0.02	 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 19
Average link density against time for 
different initial majority proportions 
1	 
0.95	 
0.9	 
0.85	 
0.8	 
0.75	 
0.7	 
0.65	 
0.6	 
0.55	 
0.5	 
0	 
5	 
10	 
15	 
20	 
25	 
30	 
35	 
40	 
45	 
50	 
55	 
60	 
65	 
70	 
75	 
80	 
85	 
90	 
95	 
100	 
mp=0.55	 
mp=0.6	 
mp=0.65	 
mp=0.7	 
mp=0.75	 
mp=0.8	 
mp=0.85	 
mp=0.9	 
mp=0.95	 
mp=1	 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 20
Conclusions 
• The particular results and insights in this 
model are not as important as the overall 
approach which… 
• …tries to get a more complex and holistic 
idea of the properties of a complex model 
• …which then might suggest simple 
hypotheses/models 
• …and thus “stage” abstraction a bit more 
gradually and carefully, being more aware 
of what is being abstracted away 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 21
Postscript: Emerging Principles 
• That evidence should not be ignored 
without a very, VERY good reason 
• That abstraction should be staged in 
gradual steps rather than “heroic” leaps 
• Be clear and explicit about your goals 
• Separate exploratory from analytic stages 
• Recognise that it is easy to fool ourselves 
and impose (wrong or limited) assumptions 
• Utilise any and all techniques that are 
applicable, but recognising their limitations 
Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 22
The End 
Claire Little: http:// 
Bruce Edmonds: http://bruce.edmonds.name 
Centre for Policy Modelling: http://cfpm.org 
Ed Fieldhouse: http:// 
Laurence Lessard-Phillips: http:// 
Institute for Social Change: http:// 
The SCID Project: http://www.scid-project.org 
These slides will be at: http://slideshare.com/BruceEdmonds 
The simulation will ‘soon’ be at: http://openabm.org as “The Voter Model”

More Related Content

Viewers also liked

Towards Institutional System Farming
Towards Institutional System FarmingTowards Institutional System Farming
Towards Institutional System FarmingBruce Edmonds
 
Modelling and Knowledge
Modelling and KnowledgeModelling and Knowledge
Modelling and KnowledgeBruce Edmonds
 
The Sociality of Context
The Sociality of ContextThe Sociality of Context
The Sociality of ContextBruce Edmonds
 
The Modelling of Context-Dependent Causal Processes A Recasting of Robert Ros...
The Modelling of Context-Dependent Causal ProcessesA Recasting of Robert Ros...The Modelling of Context-Dependent Causal ProcessesA Recasting of Robert Ros...
The Modelling of Context-Dependent Causal Processes A Recasting of Robert Ros...Bruce Edmonds
 
Winter is coming! – how to survive the coming critical storm and demonstrate ...
Winter is coming! – how to survive the coming critical storm and demonstrate ...Winter is coming! – how to survive the coming critical storm and demonstrate ...
Winter is coming! – how to survive the coming critical storm and demonstrate ...Bruce Edmonds
 
Using Agent-Based Simulation to integrate micro/qualitative evidence, macro-q...
Using Agent-Based Simulation to integrate micro/qualitative evidence, macro-q...Using Agent-Based Simulation to integrate micro/qualitative evidence, macro-q...
Using Agent-Based Simulation to integrate micro/qualitative evidence, macro-q...Bruce Edmonds
 
Computing the Sociology of Survival – how to use simulations to understand c...
Computing the Sociology of Survival – how to use simulations to understand c...Computing the Sociology of Survival – how to use simulations to understand c...
Computing the Sociology of Survival – how to use simulations to understand c...Bruce Edmonds
 
Policy Making using Modelling in a Complex world
Policy Making using Modelling in a Complex worldPolicy Making using Modelling in a Complex world
Policy Making using Modelling in a Complex worldBruce Edmonds
 
Risk-aware policy evaluation using agent-based simulation
Risk-aware policy evaluation using agent-based simulationRisk-aware policy evaluation using agent-based simulation
Risk-aware policy evaluation using agent-based simulationBruce Edmonds
 
Staged Models for Interdisciplinary Research
Staged Models for Interdisciplinary ResearchStaged Models for Interdisciplinary Research
Staged Models for Interdisciplinary ResearchBruce Edmonds
 
Simulating Superdiversity
Simulating Superdiversity Simulating Superdiversity
Simulating Superdiversity Bruce Edmonds
 
Social complexity and coupled Socio-Ecological Systems
Social complexity and coupled Socio-Ecological SystemsSocial complexity and coupled Socio-Ecological Systems
Social complexity and coupled Socio-Ecological SystemsBruce Edmonds
 
A Model of Social and Cognitive Coherence
A Model of Social and Cognitive CoherenceA Model of Social and Cognitive Coherence
A Model of Social and Cognitive CoherenceBruce Edmonds
 
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...Bruce Edmonds
 

Viewers also liked (16)

Towards Institutional System Farming
Towards Institutional System FarmingTowards Institutional System Farming
Towards Institutional System Farming
 
Modelling and Knowledge
Modelling and KnowledgeModelling and Knowledge
Modelling and Knowledge
 
The Sociality of Context
The Sociality of ContextThe Sociality of Context
The Sociality of Context
 
The Modelling of Context-Dependent Causal Processes A Recasting of Robert Ros...
The Modelling of Context-Dependent Causal ProcessesA Recasting of Robert Ros...The Modelling of Context-Dependent Causal ProcessesA Recasting of Robert Ros...
The Modelling of Context-Dependent Causal Processes A Recasting of Robert Ros...
 
Winter is coming! – how to survive the coming critical storm and demonstrate ...
Winter is coming! – how to survive the coming critical storm and demonstrate ...Winter is coming! – how to survive the coming critical storm and demonstrate ...
Winter is coming! – how to survive the coming critical storm and demonstrate ...
 
Using Agent-Based Simulation to integrate micro/qualitative evidence, macro-q...
Using Agent-Based Simulation to integrate micro/qualitative evidence, macro-q...Using Agent-Based Simulation to integrate micro/qualitative evidence, macro-q...
Using Agent-Based Simulation to integrate micro/qualitative evidence, macro-q...
 
Be ea-talk-final
Be ea-talk-finalBe ea-talk-final
Be ea-talk-final
 
Computing the Sociology of Survival – how to use simulations to understand c...
Computing the Sociology of Survival – how to use simulations to understand c...Computing the Sociology of Survival – how to use simulations to understand c...
Computing the Sociology of Survival – how to use simulations to understand c...
 
Policy Making using Modelling in a Complex world
Policy Making using Modelling in a Complex worldPolicy Making using Modelling in a Complex world
Policy Making using Modelling in a Complex world
 
A Model of Making
A Model of MakingA Model of Making
A Model of Making
 
Risk-aware policy evaluation using agent-based simulation
Risk-aware policy evaluation using agent-based simulationRisk-aware policy evaluation using agent-based simulation
Risk-aware policy evaluation using agent-based simulation
 
Staged Models for Interdisciplinary Research
Staged Models for Interdisciplinary ResearchStaged Models for Interdisciplinary Research
Staged Models for Interdisciplinary Research
 
Simulating Superdiversity
Simulating Superdiversity Simulating Superdiversity
Simulating Superdiversity
 
Social complexity and coupled Socio-Ecological Systems
Social complexity and coupled Socio-Ecological SystemsSocial complexity and coupled Socio-Ecological Systems
Social complexity and coupled Socio-Ecological Systems
 
A Model of Social and Cognitive Coherence
A Model of Social and Cognitive CoherenceA Model of Social and Cognitive Coherence
A Model of Social and Cognitive Coherence
 
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...
 

Similar to Analysing a Complex Agent-Based Model Using Data-Mining Techniques

algorithmic-decisions, fairness, machine learning, provenance, transparency
algorithmic-decisions, fairness, machine learning, provenance, transparencyalgorithmic-decisions, fairness, machine learning, provenance, transparency
algorithmic-decisions, fairness, machine learning, provenance, transparencyPaolo Missier
 
02 Network Data Collection
02 Network Data Collection02 Network Data Collection
02 Network Data Collectiondnac
 
Inverse Modeling for Cognitive Science "in the Wild"
Inverse Modeling for Cognitive Science "in the Wild"Inverse Modeling for Cognitive Science "in the Wild"
Inverse Modeling for Cognitive Science "in the Wild"Aalto University
 
Statistical Inference for development statistical model.pptx
Statistical Inference for development statistical model.pptxStatistical Inference for development statistical model.pptx
Statistical Inference for development statistical model.pptxQasimGull
 
Discovering emerging effects in Learning Networks with simulations Hendrik Dr...
Discovering emerging effects in Learning Networks with simulations Hendrik Dr...Discovering emerging effects in Learning Networks with simulations Hendrik Dr...
Discovering emerging effects in Learning Networks with simulations Hendrik Dr...Hendrik Drachsler
 
Introduction to Data and Computation: Essential capabilities for everyone in ...
Introduction to Data and Computation: Essential capabilities for everyone in ...Introduction to Data and Computation: Essential capabilities for everyone in ...
Introduction to Data and Computation: Essential capabilities for everyone in ...Kim Flintoff
 
Introduction to statistics 2013
Introduction to statistics 2013Introduction to statistics 2013
Introduction to statistics 2013Mohammad Ihmeidan
 
Sampling Design in Applied Marketing Research
Sampling Design in Applied Marketing ResearchSampling Design in Applied Marketing Research
Sampling Design in Applied Marketing ResearchKelly Page
 
DATA ANALYTICS ASSIGNMENT.pptx
DATA ANALYTICS ASSIGNMENT.pptxDATA ANALYTICS ASSIGNMENT.pptx
DATA ANALYTICS ASSIGNMENT.pptxSamirkumar497189
 
DATA MINING.doc
DATA MINING.docDATA MINING.doc
DATA MINING.docbutest
 
Mukha ng research methodology
Mukha ng research methodologyMukha ng research methodology
Mukha ng research methodologyGAMALI Roper
 
probability versusnonprobsampling
 probability versusnonprobsampling probability versusnonprobsampling
probability versusnonprobsamplingWander Guerra
 
G. Barcaroli, The use of machine learning in official statistics
G. Barcaroli, The use of machine learning in official statisticsG. Barcaroli, The use of machine learning in official statistics
G. Barcaroli, The use of machine learning in official statisticsIstituto nazionale di statistica
 

Similar to Analysing a Complex Agent-Based Model Using Data-Mining Techniques (20)

SAMPLING.pptx
SAMPLING.pptxSAMPLING.pptx
SAMPLING.pptx
 
Data collection, Data Integration, Data Understanding e Data Cleaning & Prepa...
Data collection, Data Integration, Data Understanding e Data Cleaning & Prepa...Data collection, Data Integration, Data Understanding e Data Cleaning & Prepa...
Data collection, Data Integration, Data Understanding e Data Cleaning & Prepa...
 
Sampaling
SampalingSampaling
Sampaling
 
algorithmic-decisions, fairness, machine learning, provenance, transparency
algorithmic-decisions, fairness, machine learning, provenance, transparencyalgorithmic-decisions, fairness, machine learning, provenance, transparency
algorithmic-decisions, fairness, machine learning, provenance, transparency
 
02 Network Data Collection
02 Network Data Collection02 Network Data Collection
02 Network Data Collection
 
02 Network Data Collection (2016)
02 Network Data Collection (2016)02 Network Data Collection (2016)
02 Network Data Collection (2016)
 
Inverse Modeling for Cognitive Science "in the Wild"
Inverse Modeling for Cognitive Science "in the Wild"Inverse Modeling for Cognitive Science "in the Wild"
Inverse Modeling for Cognitive Science "in the Wild"
 
Data mining
Data miningData mining
Data mining
 
Statistical Inference for development statistical model.pptx
Statistical Inference for development statistical model.pptxStatistical Inference for development statistical model.pptx
Statistical Inference for development statistical model.pptx
 
Data mining BY Zubair Yaseen
Data mining BY Zubair YaseenData mining BY Zubair Yaseen
Data mining BY Zubair Yaseen
 
Discovering emerging effects in Learning Networks with simulations Hendrik Dr...
Discovering emerging effects in Learning Networks with simulations Hendrik Dr...Discovering emerging effects in Learning Networks with simulations Hendrik Dr...
Discovering emerging effects in Learning Networks with simulations Hendrik Dr...
 
Introduction to Data and Computation: Essential capabilities for everyone in ...
Introduction to Data and Computation: Essential capabilities for everyone in ...Introduction to Data and Computation: Essential capabilities for everyone in ...
Introduction to Data and Computation: Essential capabilities for everyone in ...
 
Introduction to statistics 2013
Introduction to statistics 2013Introduction to statistics 2013
Introduction to statistics 2013
 
Sampling Design in Applied Marketing Research
Sampling Design in Applied Marketing ResearchSampling Design in Applied Marketing Research
Sampling Design in Applied Marketing Research
 
2018 Modern Math Workshop - Nonparametric Regression and Classification for M...
2018 Modern Math Workshop - Nonparametric Regression and Classification for M...2018 Modern Math Workshop - Nonparametric Regression and Classification for M...
2018 Modern Math Workshop - Nonparametric Regression and Classification for M...
 
DATA ANALYTICS ASSIGNMENT.pptx
DATA ANALYTICS ASSIGNMENT.pptxDATA ANALYTICS ASSIGNMENT.pptx
DATA ANALYTICS ASSIGNMENT.pptx
 
DATA MINING.doc
DATA MINING.docDATA MINING.doc
DATA MINING.doc
 
Mukha ng research methodology
Mukha ng research methodologyMukha ng research methodology
Mukha ng research methodology
 
probability versusnonprobsampling
 probability versusnonprobsampling probability versusnonprobsampling
probability versusnonprobsampling
 
G. Barcaroli, The use of machine learning in official statistics
G. Barcaroli, The use of machine learning in official statisticsG. Barcaroli, The use of machine learning in official statistics
G. Barcaroli, The use of machine learning in official statistics
 

More from Bruce Edmonds

Staging Model Abstraction – an example about political participation
Staging Model Abstraction – an example about political participationStaging Model Abstraction – an example about political participation
Staging Model Abstraction – an example about political participationBruce Edmonds
 
Modelling Pitfalls - extra resources
Modelling Pitfalls - extra resourcesModelling Pitfalls - extra resources
Modelling Pitfalls - extra resourcesBruce Edmonds
 
Modelling Pitfalls - introduction and some cases
Modelling Pitfalls - introduction and some casesModelling Pitfalls - introduction and some cases
Modelling Pitfalls - introduction and some casesBruce Edmonds
 
The evolution of empirical ABMs
The evolution of empirical ABMsThe evolution of empirical ABMs
The evolution of empirical ABMsBruce Edmonds
 
Mixing fat data, simulation and policy - what could possibly go wrong?
Mixing fat data, simulation and policy - what could possibly go wrong?Mixing fat data, simulation and policy - what could possibly go wrong?
Mixing fat data, simulation and policy - what could possibly go wrong?Bruce Edmonds
 
Using agent-based simulation for socio-ecological uncertainty analysis
Using agent-based simulation for socio-ecological uncertainty analysisUsing agent-based simulation for socio-ecological uncertainty analysis
Using agent-based simulation for socio-ecological uncertainty analysisBruce Edmonds
 
Finding out what could go wrong before it does – Modelling Risk and Uncertainty
Finding out what could go wrong before it does – Modelling Risk and UncertaintyFinding out what could go wrong before it does – Modelling Risk and Uncertainty
Finding out what could go wrong before it does – Modelling Risk and UncertaintyBruce Edmonds
 
How social simulation could help social science deal with context
How social simulation could help social science deal with contextHow social simulation could help social science deal with context
How social simulation could help social science deal with contextBruce Edmonds
 
Agent-based modelling, laboratory experiments, and observation in the wild
Agent-based modelling,laboratory experiments,and observation in the wildAgent-based modelling,laboratory experiments,and observation in the wild
Agent-based modelling, laboratory experiments, and observation in the wildBruce Edmonds
 
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...
Culture trumps ethnicity!– Intra-generational cultural evolution and ethnoce...Culture trumps ethnicity!– Intra-generational cultural evolution and ethnoce...
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...Bruce Edmonds
 
An Introduction to Agent-Based Modelling
An Introduction to Agent-Based ModellingAn Introduction to Agent-Based Modelling
An Introduction to Agent-Based ModellingBruce Edmonds
 
Mixing ABM and policy...what could possibly go wrong?
Mixing ABM and policy...what could possibly go wrong?Mixing ABM and policy...what could possibly go wrong?
Mixing ABM and policy...what could possibly go wrong?Bruce Edmonds
 
Different Modelling Purposes - an 'anit-theoretical' approach
Different Modelling Purposes - an 'anit-theoretical' approachDifferent Modelling Purposes - an 'anit-theoretical' approach
Different Modelling Purposes - an 'anit-theoretical' approachBruce Edmonds
 
Socio-Ecological Simulation - a risk-assessment approach
Socio-Ecological Simulation - a risk-assessment approachSocio-Ecological Simulation - a risk-assessment approach
Socio-Ecological Simulation - a risk-assessment approachBruce Edmonds
 
A Simple Model of Group Commoning
A Simple Model of Group CommoningA Simple Model of Group Commoning
A Simple Model of Group CommoningBruce Edmonds
 
6 Modelling Purposes
6 Modelling Purposes6 Modelling Purposes
6 Modelling PurposesBruce Edmonds
 
Are Mixed-Methods Just a Fudge? The Dangers and Prospects for Integrating Qu...
Are Mixed-Methods Just a Fudge? The Dangers and Prospects for Integrating Qu...Are Mixed-Methods Just a Fudge? The Dangers and Prospects for Integrating Qu...
Are Mixed-Methods Just a Fudge? The Dangers and Prospects for Integrating Qu...Bruce Edmonds
 
The Post-Truth Drift in Social Simulation
The Post-Truth Drift in Social SimulationThe Post-Truth Drift in Social Simulation
The Post-Truth Drift in Social SimulationBruce Edmonds
 
Drilling down below opinions: how co-evolving beliefs and social structure mi...
Drilling down below opinions: how co-evolving beliefs and social structure mi...Drilling down below opinions: how co-evolving beliefs and social structure mi...
Drilling down below opinions: how co-evolving beliefs and social structure mi...Bruce Edmonds
 

More from Bruce Edmonds (20)

Staging Model Abstraction – an example about political participation
Staging Model Abstraction – an example about political participationStaging Model Abstraction – an example about political participation
Staging Model Abstraction – an example about political participation
 
Modelling Pitfalls - extra resources
Modelling Pitfalls - extra resourcesModelling Pitfalls - extra resources
Modelling Pitfalls - extra resources
 
Modelling Pitfalls - introduction and some cases
Modelling Pitfalls - introduction and some casesModelling Pitfalls - introduction and some cases
Modelling Pitfalls - introduction and some cases
 
The evolution of empirical ABMs
The evolution of empirical ABMsThe evolution of empirical ABMs
The evolution of empirical ABMs
 
Mixing fat data, simulation and policy - what could possibly go wrong?
Mixing fat data, simulation and policy - what could possibly go wrong?Mixing fat data, simulation and policy - what could possibly go wrong?
Mixing fat data, simulation and policy - what could possibly go wrong?
 
Social Context
Social ContextSocial Context
Social Context
 
Using agent-based simulation for socio-ecological uncertainty analysis
Using agent-based simulation for socio-ecological uncertainty analysisUsing agent-based simulation for socio-ecological uncertainty analysis
Using agent-based simulation for socio-ecological uncertainty analysis
 
Finding out what could go wrong before it does – Modelling Risk and Uncertainty
Finding out what could go wrong before it does – Modelling Risk and UncertaintyFinding out what could go wrong before it does – Modelling Risk and Uncertainty
Finding out what could go wrong before it does – Modelling Risk and Uncertainty
 
How social simulation could help social science deal with context
How social simulation could help social science deal with contextHow social simulation could help social science deal with context
How social simulation could help social science deal with context
 
Agent-based modelling, laboratory experiments, and observation in the wild
Agent-based modelling,laboratory experiments,and observation in the wildAgent-based modelling,laboratory experiments,and observation in the wild
Agent-based modelling, laboratory experiments, and observation in the wild
 
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...
Culture trumps ethnicity!– Intra-generational cultural evolution and ethnoce...Culture trumps ethnicity!– Intra-generational cultural evolution and ethnoce...
Culture trumps ethnicity! – Intra-generational cultural evolution and ethnoce...
 
An Introduction to Agent-Based Modelling
An Introduction to Agent-Based ModellingAn Introduction to Agent-Based Modelling
An Introduction to Agent-Based Modelling
 
Mixing ABM and policy...what could possibly go wrong?
Mixing ABM and policy...what could possibly go wrong?Mixing ABM and policy...what could possibly go wrong?
Mixing ABM and policy...what could possibly go wrong?
 
Different Modelling Purposes - an 'anit-theoretical' approach
Different Modelling Purposes - an 'anit-theoretical' approachDifferent Modelling Purposes - an 'anit-theoretical' approach
Different Modelling Purposes - an 'anit-theoretical' approach
 
Socio-Ecological Simulation - a risk-assessment approach
Socio-Ecological Simulation - a risk-assessment approachSocio-Ecological Simulation - a risk-assessment approach
Socio-Ecological Simulation - a risk-assessment approach
 
A Simple Model of Group Commoning
A Simple Model of Group CommoningA Simple Model of Group Commoning
A Simple Model of Group Commoning
 
6 Modelling Purposes
6 Modelling Purposes6 Modelling Purposes
6 Modelling Purposes
 
Are Mixed-Methods Just a Fudge? The Dangers and Prospects for Integrating Qu...
Are Mixed-Methods Just a Fudge? The Dangers and Prospects for Integrating Qu...Are Mixed-Methods Just a Fudge? The Dangers and Prospects for Integrating Qu...
Are Mixed-Methods Just a Fudge? The Dangers and Prospects for Integrating Qu...
 
The Post-Truth Drift in Social Simulation
The Post-Truth Drift in Social SimulationThe Post-Truth Drift in Social Simulation
The Post-Truth Drift in Social Simulation
 
Drilling down below opinions: how co-evolving beliefs and social structure mi...
Drilling down below opinions: how co-evolving beliefs and social structure mi...Drilling down below opinions: how co-evolving beliefs and social structure mi...
Drilling down below opinions: how co-evolving beliefs and social structure mi...
 

Recently uploaded

Separation of Lanthanides/ Lanthanides and Actinides
Separation of Lanthanides/ Lanthanides and ActinidesSeparation of Lanthanides/ Lanthanides and Actinides
Separation of Lanthanides/ Lanthanides and ActinidesFatimaKhan178732
 
Organic Name Reactions for the students and aspirants of Chemistry12th.pptx
Organic Name Reactions  for the students and aspirants of Chemistry12th.pptxOrganic Name Reactions  for the students and aspirants of Chemistry12th.pptx
Organic Name Reactions for the students and aspirants of Chemistry12th.pptxVS Mahajan Coaching Centre
 
The Most Excellent Way | 1 Corinthians 13
The Most Excellent Way | 1 Corinthians 13The Most Excellent Way | 1 Corinthians 13
The Most Excellent Way | 1 Corinthians 13Steve Thomason
 
Employee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptxEmployee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptxNirmalaLoungPoorunde1
 
Mastering the Unannounced Regulatory Inspection
Mastering the Unannounced Regulatory InspectionMastering the Unannounced Regulatory Inspection
Mastering the Unannounced Regulatory InspectionSafetyChain Software
 
microwave assisted reaction. General introduction
microwave assisted reaction. General introductionmicrowave assisted reaction. General introduction
microwave assisted reaction. General introductionMaksud Ahmed
 
_Math 4-Q4 Week 5.pptx Steps in Collecting Data
_Math 4-Q4 Week 5.pptx Steps in Collecting Data_Math 4-Q4 Week 5.pptx Steps in Collecting Data
_Math 4-Q4 Week 5.pptx Steps in Collecting DataJhengPantaleon
 
Call Girls in Dwarka Mor Delhi Contact Us 9654467111
Call Girls in Dwarka Mor Delhi Contact Us 9654467111Call Girls in Dwarka Mor Delhi Contact Us 9654467111
Call Girls in Dwarka Mor Delhi Contact Us 9654467111Sapana Sha
 
URLs and Routing in the Odoo 17 Website App
URLs and Routing in the Odoo 17 Website AppURLs and Routing in the Odoo 17 Website App
URLs and Routing in the Odoo 17 Website AppCeline George
 
KSHARA STURA .pptx---KSHARA KARMA THERAPY (CAUSTIC THERAPY)————IMP.OF KSHARA ...
KSHARA STURA .pptx---KSHARA KARMA THERAPY (CAUSTIC THERAPY)————IMP.OF KSHARA ...KSHARA STURA .pptx---KSHARA KARMA THERAPY (CAUSTIC THERAPY)————IMP.OF KSHARA ...
KSHARA STURA .pptx---KSHARA KARMA THERAPY (CAUSTIC THERAPY)————IMP.OF KSHARA ...M56BOOKSTORE PRODUCT/SERVICE
 
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17Celine George
 
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...Marc Dusseiller Dusjagr
 
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPTECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPTiammrhaywood
 
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...Krashi Coaching
 
Presiding Officer Training module 2024 lok sabha elections
Presiding Officer Training module 2024 lok sabha electionsPresiding Officer Training module 2024 lok sabha elections
Presiding Officer Training module 2024 lok sabha electionsanshu789521
 
Concept of Vouching. B.Com(Hons) /B.Compdf
Concept of Vouching. B.Com(Hons) /B.CompdfConcept of Vouching. B.Com(Hons) /B.Compdf
Concept of Vouching. B.Com(Hons) /B.CompdfUmakantAnnand
 
A Critique of the Proposed National Education Policy Reform
A Critique of the Proposed National Education Policy ReformA Critique of the Proposed National Education Policy Reform
A Critique of the Proposed National Education Policy ReformChameera Dedduwage
 

Recently uploaded (20)

Separation of Lanthanides/ Lanthanides and Actinides
Separation of Lanthanides/ Lanthanides and ActinidesSeparation of Lanthanides/ Lanthanides and Actinides
Separation of Lanthanides/ Lanthanides and Actinides
 
9953330565 Low Rate Call Girls In Rohini Delhi NCR
9953330565 Low Rate Call Girls In Rohini  Delhi NCR9953330565 Low Rate Call Girls In Rohini  Delhi NCR
9953330565 Low Rate Call Girls In Rohini Delhi NCR
 
Organic Name Reactions for the students and aspirants of Chemistry12th.pptx
Organic Name Reactions  for the students and aspirants of Chemistry12th.pptxOrganic Name Reactions  for the students and aspirants of Chemistry12th.pptx
Organic Name Reactions for the students and aspirants of Chemistry12th.pptx
 
The Most Excellent Way | 1 Corinthians 13
The Most Excellent Way | 1 Corinthians 13The Most Excellent Way | 1 Corinthians 13
The Most Excellent Way | 1 Corinthians 13
 
Employee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptxEmployee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptx
 
Mastering the Unannounced Regulatory Inspection
Mastering the Unannounced Regulatory InspectionMastering the Unannounced Regulatory Inspection
Mastering the Unannounced Regulatory Inspection
 
microwave assisted reaction. General introduction
microwave assisted reaction. General introductionmicrowave assisted reaction. General introduction
microwave assisted reaction. General introduction
 
_Math 4-Q4 Week 5.pptx Steps in Collecting Data
_Math 4-Q4 Week 5.pptx Steps in Collecting Data_Math 4-Q4 Week 5.pptx Steps in Collecting Data
_Math 4-Q4 Week 5.pptx Steps in Collecting Data
 
Call Girls in Dwarka Mor Delhi Contact Us 9654467111
Call Girls in Dwarka Mor Delhi Contact Us 9654467111Call Girls in Dwarka Mor Delhi Contact Us 9654467111
Call Girls in Dwarka Mor Delhi Contact Us 9654467111
 
URLs and Routing in the Odoo 17 Website App
URLs and Routing in the Odoo 17 Website AppURLs and Routing in the Odoo 17 Website App
URLs and Routing in the Odoo 17 Website App
 
KSHARA STURA .pptx---KSHARA KARMA THERAPY (CAUSTIC THERAPY)————IMP.OF KSHARA ...
KSHARA STURA .pptx---KSHARA KARMA THERAPY (CAUSTIC THERAPY)————IMP.OF KSHARA ...KSHARA STURA .pptx---KSHARA KARMA THERAPY (CAUSTIC THERAPY)————IMP.OF KSHARA ...
KSHARA STURA .pptx---KSHARA KARMA THERAPY (CAUSTIC THERAPY)————IMP.OF KSHARA ...
 
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
 
Staff of Color (SOC) Retention Efforts DDSD
Staff of Color (SOC) Retention Efforts DDSDStaff of Color (SOC) Retention Efforts DDSD
Staff of Color (SOC) Retention Efforts DDSD
 
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
 
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPTECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
 
Model Call Girl in Tilak Nagar Delhi reach out to us at 🔝9953056974🔝
Model Call Girl in Tilak Nagar Delhi reach out to us at 🔝9953056974🔝Model Call Girl in Tilak Nagar Delhi reach out to us at 🔝9953056974🔝
Model Call Girl in Tilak Nagar Delhi reach out to us at 🔝9953056974🔝
 
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
 
Presiding Officer Training module 2024 lok sabha elections
Presiding Officer Training module 2024 lok sabha electionsPresiding Officer Training module 2024 lok sabha elections
Presiding Officer Training module 2024 lok sabha elections
 
Concept of Vouching. B.Com(Hons) /B.Compdf
Concept of Vouching. B.Com(Hons) /B.CompdfConcept of Vouching. B.Com(Hons) /B.Compdf
Concept of Vouching. B.Com(Hons) /B.Compdf
 
A Critique of the Proposed National Education Policy Reform
A Critique of the Proposed National Education Policy ReformA Critique of the Proposed National Education Policy Reform
A Critique of the Proposed National Education Policy Reform
 

Analysing a Complex Agent-Based Model Using Data-Mining Techniques

  • 1. Analysing a Complex Agent-Based Model Using Data-Mining Techniques Claire Little, Bruce Edmonds Centre for Policy Modelling Manchester Metropolitan University Ed Fieldhouse, Laurence Lessard-Phillips Institute for Social Change University of Manchester
  • 2. An “Inconvenient Truth” • That the universe is not arranged for our benefit (as researchers studying it) • in other words, that assumptions such as the following are likely to be wrong: – Our planet is the centre of the universe – Risky events follow a normal distribution – Humans act as if they followed a simple algorithm – Society can be understood using simple, universal prinicples – etc. etc. • In particular, the assumption that there will always (if you look hard enough) be models that are: – (a) simple enough for us to understand and – (b) adequate to what we want to model • …is wrong Complexity and Context-Dependency, Bruce Edmonds, ECCS, Lisbon, Sept 2010. slide-2
  • 3. The Alternative • Thus consider the alternative, more realistic, situation where one is facing some phenomena where any model that is adequate (w.r.t. our goals) will be too complex for us to completely understand • Instead of indulging in wishful thinking this paper looks at ways forward under complexity • In other words, if we have a simulation model that is too complex to completely understand, how can we obtain some useful understanding of its properties… • …and hence use it to leverage some understanding/control over the target phenomena Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 3
  • 4. The Broad Idea 1. Make relatively complex simulations based on available evidence (a “KIDS” approach) 2. Analyse this simulation in a number of ways, including… 3. Data mining output data in a more holistic manner over a broad “space” of settings 4. Look for patterns in the data that suggest (maybe context-dependent) hypotheses 5. These are candidates for simpler (but maybe partial) models of the simulation Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 4
  • 5. The Model • To explore the complex mix of factors, structures and processes that affect whether people vote • An agent-based model, with demographics and dynamic social networks • Was formulated using a mixture of qualitative, survey data and others’ expert opinion/results Class Age Activities Ethnicity Etc. Level-of-Political-Interest A Household Discuss-politics-with person-23 blue expert=false neighbour-network year=10 month=3 Lots-family-discussions year=10 month=2 Etc. Memory An Agent’s Memory of Events Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 5
  • 6. Overall Structure of Model Underlying data about population composition Demographics of people in households Social network formation and maintenance (homophily) Influence via social networks • Political discussions Voting Behaviour Input Output
  • 7. Technique • Instead of initiating ‘thin’ analyses of the simulation behaviour (e.g. 1 or 2D parameter sweeps/correlation models against a few key output measures) • To sample a multi-dimensional space of settings and cluster on a multi-dimensional space of output indicators (in this case 9 parameter x 13 output measures) • Look at the patterns between clusters for indications as to hypotheses of behaviour • Then test these with targeted simulation experiments Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 7
  • 8. More Holistic but Less Detailed 1. Many Runs, Randomly Sampling Parameter Values Multi-Dimensional Space of Parameter Settings 2. Analyse data set of result measures using data mining 3. Look for patterns that you might then check in a more systematic manner Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 8
  • 9. Parameter ranges 3862 independent runs with parameters sampled from the following, uniform distributions: • density: [0.65, 0.95] • drop-activity-prob: [0.05, 0.15] • drop-friend-prob: [0, 0.01] • emmigration-rate: [0 ,0.03] • immigration-rate: [0, 0.02] • int-immigration-rate: [0, 0.02] • majority-prop: [0.55, 1] • prob-move-near: [0, 1] • prob-partner: [0.01, 0.03] Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 9
  • 10. For each of these runs… • Measure many different indicators of the outputs (say at the end of the simulation) including: – Pop.size – population size – Av.age – average age – Av.adfriends – average number of friends (adults only) – Prop.maj – proportion of the majority population – Prop.adult – proportion that is adult – Prop.1stgen – proportion that are 1st generation immigrant – av.clust – average proportion of friends who are friends – av.sim.hh – average similarity within households – av.sim.fr – average similarity between friends – ncvs.ac – number of conversations over activity links – ncvs.sc – number of conversations over “school” links – Prop. Adults with highest level of political interest Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 10
  • 11. Dendrogram of hierarchical clustering of simulations Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 11
  • 12. A heatmap of the hierarchical clustering Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 12
  • 13. The within group sum of squares against the number of clusters for 10 randomly initialised runs using k-means Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 13
  • 14. Clustergram of PCA-weighted mean of k-mean clusters vs. number of clusters Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 14
  • 15. Centroid plot against the first two discriminant functions showing the 3 clusters Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 15
  • 16. Details of the centroids of the 3 k-means clusters Attribute Cluster 1 (543 records) Cluster 2 (1333 records) Cluster 3 ( 1986 records) Pop.size 100 557 1750 Av.age 76 58 55 Av.adfriends 0.73 1.36 1.82 Prop.maj 74% 67% 65% Prop.adult 99% 94% 93.5% Prop.1stgen 8% 13% 14% av.clust 0.97 0.84 0.70 av.sim.hh 2.45 3.53 3.74 av.sim.fr 2.82 3.70 3.33 Rate ncvs.ac 1.3% 1.3% 0.0% Rate ncvs.sc 0.45% 0.20% 0.13% Prop. Adults 0.97% 1.6% 1.7% involved Within cluster sum of squares 6748.243 11288.460 7407.591 Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 16
  • 17. the 3 clusters against the parameters: emigration rate, immigration rate, internal immigration rate Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 17
  • 18. Pop Size Av. Age Multi-Dimensional Scatter Graphs Av Sim Hh Av Sim Fr Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 18
  • 19. Average proportion of similar friends against time for different immigration rates 0.95 0.9 0.85 0.8 0.75 0.7 0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 0 0.0025 0.005 0.0075 0.01 0.0125 0.015 0.0175 0.02 Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 19
  • 20. Average link density against time for different initial majority proportions 1 0.95 0.9 0.85 0.8 0.75 0.7 0.65 0.6 0.55 0.5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 mp=0.55 mp=0.6 mp=0.65 mp=0.7 mp=0.75 mp=0.8 mp=0.85 mp=0.9 mp=0.95 mp=1 Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 20
  • 21. Conclusions • The particular results and insights in this model are not as important as the overall approach which… • …tries to get a more complex and holistic idea of the properties of a complex model • …which then might suggest simple hypotheses/models • …and thus “stage” abstraction a bit more gradually and carefully, being more aware of what is being abstracted away Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 21
  • 22. Postscript: Emerging Principles • That evidence should not be ignored without a very, VERY good reason • That abstraction should be staged in gradual steps rather than “heroic” leaps • Be clear and explicit about your goals • Separate exploratory from analytic stages • Recognise that it is easy to fool ourselves and impose (wrong or limited) assumptions • Utilise any and all techniques that are applicable, but recognising their limitations Analysing a Complex Agent-Based Model Using Data-Mining Techniques. Little & Edmonds, SSC 2014, Barcelona, 22
  • 23. The End Claire Little: http:// Bruce Edmonds: http://bruce.edmonds.name Centre for Policy Modelling: http://cfpm.org Ed Fieldhouse: http:// Laurence Lessard-Phillips: http:// Institute for Social Change: http:// The SCID Project: http://www.scid-project.org These slides will be at: http://slideshare.com/BruceEdmonds The simulation will ‘soon’ be at: http://openabm.org as “The Voter Model”