SlideShare a Scribd company logo
Center for Financial Studies at the Goethe University
PhD Mini-course
Frankfurt, 25 January 2013




Financial Networks

VI. Correlation Networks




                                                        Dr. Kimmo Soramäki
                                                        Founder and CEO
                                                        FNA, www.fna.fi
Agenda

V. Inferring Links
•   Prices and returns
•   Controlling for common factors
•   Correlation and dependence
•   Significant correlations
•   Multiple Comparisons


VI. Correlation Networks
•   Distance and Hierarchical Clustering
•   Minimum Spanning Tree & PMFG
•   Other filtering
•   Layout algorithms
                                           2
Hierarchical structure in financial markets




                                              3
Minimum Spanning Tree
A spanning tree of a graph is a subgraph that:
1. is a tree and
2. connects all the nodes together

Length of a tree is the sum of its links. Minimum spanning tree (MST) is a spanning
tree with shortest length.

MST reflects the hierarchical structure of the correlation matrix
MST and Hierarchical Structure




Source: R.N. Mantegna (1999). Hierarchical structure in nancial markets,
Eur. Phys. J. B 11, 193-197                                                5
36


Single Linkage Clustering

• A method for hierarchical clustering
• Clusters based on similarity or distance
• SLINK algorithm




R. Sibson (1973). SLINK: an optimally efficient algorithm for the single-link cluster
method. The Computer Journal (British Computer Society) 16 (1): 30–34.                  6
Example
# build network from correlations
buildbycorrelationd -file daxreturns-2011-recon.csv -missing Alert -preserve
false

# calculate distance
corrdistance -p correlation -method gower

# calculate single linkage clistering
slink -p corrdistance

# create heatmaps
heatmap -sortv vertex_id -p correlation -symmetric true -cellsizedefault 13 -
transition 0 -cellhover correlation -palette darkblue-lightgray-darkred -
colordomain (-1)-1 -saveas daxheat-slink-Y

                                                                                7
Unordered, Principal   Ordered by Cluster, Principal
Component Removed      Component Removed

                                                       8
Radial tree -layout
• Calculates coordinates for radial layout as presented in
  Bachmaier, Brandes and Schlieper (2005)

• The layout allows definition of each arc length

• Specific parameters of command radialtreeviz:
     –   Arc length property (-p) : Arc property defining arc length. Optional.
     –   Root vertex (-rootvertex) : Id of root vertex. The root vertex is placed in the middle of the
         screen. Due to the repositioning of the tree, nodes may be placed outside the canvas in other
         than the first network. Optional.
     –   Optimal rotation (-rotation) : Rotates layout to minimize sum of vertex distances between
         subsequent networks. Optional. By default 'false'.
     –   Scaling (-scale) : Scale of visualization: value/pixel.




Christian Bachmaier, Ulrik Brandes, and Barbara Schlieper (2005). Drawing Phylogenetic
Trees. Department of Computer & Information Science, University of
                                                                                                         9
Konstanz, Germany
Putting it all together
# build network from correlations
buildbycorrelationd -file daxreturns-2011.csv -missing Alert -savestdev -savereturns -preserve false

# calculate distance
corrdistance -p correlation -method gower

# calculate single linkage clistering
minst -p corrdistance

# drop arcs not in MST
dropa -e minst=false

# calculate absolute correlation
calcap -e 1-abs(correlation) -saveas vizdistance

# create heatmaps
radialtreeviz -p vizdistance -vlabel vertex_id -vsize stdev -transition 3000 -ahover correlation -saveas
daxviz-MST

                                                                                                       10
Asset Trees
              Size of node reflects volatility
              (variance) of returns

              Links between nodes reflect
              'backbone' correlations

              - short link = high correlation
              - long link = low correlation




                                                 11
Circle Tree -visualization
• Calculates coordinates for circle
  tree layout as presented in
  Bachmaier, Brandes and
  Schlieper (2005)



• As before but instead of
  radialtreeviz:

circletreeviz -vlabel vertex_id -vsize
stdev -transition 3000 -ahover
correlation -saveas daxviz-MST-circle




                                         12
Planar Maximally Filtered Graph
                                                                                                         Node size scales
                                                                                                         with degree

•   A complex graph with loops and
    cliques of up to 4 elements. It can be
    drawn on a planar surface without
    link crossings.

•   MST is contained in PMFG




    M. Tumminello, T. Ast, T. Di Matteo and R. N. Mantegna (2005). A Tool for Filtering Information in
    Complex Systems. PNAS vol. 102 no. 30 pp. 10421–10426
                                                                                                                   13
PMFG -command
# build network from correlations
buildbycorrelationd -file daxreturns-2011.csv -missing Alert -savestdev -savereturns -preserve false

# calculate distance
corrdistance -p correlation -method gower

# calculate single linkage clistering
pmfg -p corrdistance

# drop arcs not in MST
dropa -e pmfg=false

# calculate 1-absolute correlation
calcap -e abs(correlation) -saveas vizdistance

# calculate degree
degree

# create heatmaps
frviz -vlabel vertex_id -vsize stdev -atransparency vizdistance -ahover correlation -transition 3000 -ahover correlation -
arrows false -saveas daxviz-PMFG

                                                                                                                        14
Partial Correlation
• Measures the degree of association between two random variables

• What is the direct relationship between Adidas and
  Allianz, controlling for BASF, BAYER, ... ?

• We build regression models for Adidas and Allianz and look at the
  correlation of their model residuals (i.e. wgat left unexplained by the
  other factors) -> Partial correltation




                                                                        15
Example
# build network from correlations
buildbypartialcorrelationd -file daxreturns-2011.csv -missing Alert -
savestdev -preserve false

# show as heatmap
heatmap -sortv vertex_id -p partial_correlation -symmetric true -
cellsizedefault 13 -transition 0 -cellhover partial_correlation -palette
darkblue-lightgray-darkred -colordomain (-1)-1 -saveas daxheat-
partial-Y




                                                                           16
Correlations   Partial Correlations


                                      17
NETS
•   Network Estimation for Time-
    Series

•   Forthcoming paper by Barigozzi
    and Brownlees

•   Estimates an unknown network
    structure from multivariate data

•   Captures both comtemporenous
    and serial dependence (partial
    correlations and lead/lag effects)



                                         18
Correlation filtering                            PMFG



Balance between too much and too little
information

One of many methods to create networks
from correlation/distance matrices
   – PMFGs, Partial Correlation
     Networks, Influence Networks, Granger   Influence Network
     Causality, NETS, etc.


New graph, information-theory, economics
& statistics -based models are being
actively developed



                                                                 19
Sammon’s Projection
Proposed by John W. Sammon in IEEE Transactions on Computers 18: 401–409
(1969)

A nonlinear projection method to map a
high dimensional space onto a space of
lower dimensionality. Example:

                                                                  Iris Setosa




                                                                Iris Versicolor




                                                                 Iris Virginica
                                                                         20
Example
# build network from correlations
buildbycorrelationd -file daxreturns-2011.csv -missing Alert -savestdev -savereturns -
preserve false

# calculate distance
corrdistance -p correlation -method gower

# Calculate sammonlayout
sammonlayouta -p corrdistance -saveerror true

# Sum up error
sumaforv -p error -saveas error

# create heatmaps
sammonaviz -p corrdistance -vlabel vertex_id -vsize error -transition 3000 -ahover error
-saveas daxviz-Sammon-Y

                                                                                         21
Node size reflects
error in layout
Tutorials
•   Tutorial 1 – Loading Networks into FNA
•   Tutorial 2 – Managing Data in FNA
•   Tutorial 3 – Network Summary Measures
•   Tutorial 4 – Centrality Measures
•   Tutorial 5 – Connectedness and Components
•   Tutorial 6 – Network Visualization
•   Tutorial 7 – Correlation Networks
•   Tutorial 8 – Payment System Simulations
•   Tutorial 9 – Analyzing Cross-Border Banking Exposures




                                                            23
Blog, Library and Demos at www.fna.fi




Dr. Kimmo Soramäki
kimmo@soramaki.net
Twitter: soramaki

More Related Content

What's hot

Graph Signal Processing for Machine Learning A Review and New Perspectives - ...
Graph Signal Processing for Machine Learning A Review and New Perspectives - ...Graph Signal Processing for Machine Learning A Review and New Perspectives - ...
Graph Signal Processing for Machine Learning A Review and New Perspectives - ...
lauratoni4
 
Csr2011 june16 11_30_georgiadis
Csr2011 june16 11_30_georgiadisCsr2011 june16 11_30_georgiadis
Csr2011 june16 11_30_georgiadis
CSR2011
 

What's hot (11)

Massive parallelism with gpus for centrality ranking in complex networks
Massive parallelism with gpus for centrality ranking in complex networksMassive parallelism with gpus for centrality ranking in complex networks
Massive parallelism with gpus for centrality ranking in complex networks
 
A few approaches to cope with crash/byzantine failures in communication netwo...
A few approaches to cope with crash/byzantine failures in communication netwo...A few approaches to cope with crash/byzantine failures in communication netwo...
A few approaches to cope with crash/byzantine failures in communication netwo...
 
Graph Signal Processing for Machine Learning A Review and New Perspectives - ...
Graph Signal Processing for Machine Learning A Review and New Perspectives - ...Graph Signal Processing for Machine Learning A Review and New Perspectives - ...
Graph Signal Processing for Machine Learning A Review and New Perspectives - ...
 
Csr2011 june16 11_30_georgiadis
Csr2011 june16 11_30_georgiadisCsr2011 june16 11_30_georgiadis
Csr2011 june16 11_30_georgiadis
 
1801.06434
1801.064341801.06434
1801.06434
 
Ba36317323
Ba36317323Ba36317323
Ba36317323
 
A genetic algorithm for constructing broadcast trees with cost and delay cons...
A genetic algorithm for constructing broadcast trees with cost and delay cons...A genetic algorithm for constructing broadcast trees with cost and delay cons...
A genetic algorithm for constructing broadcast trees with cost and delay cons...
 
CCNxCon2012: Session 5: CCN Location Sharing System
CCNxCon2012: Session 5: CCN Location Sharing SystemCCNxCon2012: Session 5: CCN Location Sharing System
CCNxCon2012: Session 5: CCN Location Sharing System
 
A New Approach to Linear Estimation Problem in Multiuser Massive MIMO Systems
A New Approach to Linear Estimation Problem in Multiuser Massive MIMO SystemsA New Approach to Linear Estimation Problem in Multiuser Massive MIMO Systems
A New Approach to Linear Estimation Problem in Multiuser Massive MIMO Systems
 
The Effect of Network Topology on Geographic Routing Performance in Localized...
The Effect of Network Topology on Geographic Routing Performance in Localized...The Effect of Network Topology on Geographic Routing Performance in Localized...
The Effect of Network Topology on Geographic Routing Performance in Localized...
 
Effective Sparse Matrix Representation for the GPU Architectures
 Effective Sparse Matrix Representation for the GPU Architectures Effective Sparse Matrix Representation for the GPU Architectures
Effective Sparse Matrix Representation for the GPU Architectures
 

Similar to Financial Networks VI - Correlation Networks

A Survey On Secure Cooperative Bait Detection Approach For...
A Survey On Secure Cooperative Bait Detection Approach For...A Survey On Secure Cooperative Bait Detection Approach For...
A Survey On Secure Cooperative Bait Detection Approach For...
Jennifer Perry
 

Similar to Financial Networks VI - Correlation Networks (20)

A TALE of DATA PATTERN DISCOVERY IN PARALLEL
A TALE of DATA PATTERN DISCOVERY IN PARALLELA TALE of DATA PATTERN DISCOVERY IN PARALLEL
A TALE of DATA PATTERN DISCOVERY IN PARALLEL
 
High-Performance Graph Analysis and Modeling
High-Performance Graph Analysis and ModelingHigh-Performance Graph Analysis and Modeling
High-Performance Graph Analysis and Modeling
 
Financial Networks: II. Fundamentals of Network Theory and FNA
Financial Networks: II. Fundamentals of Network Theory and FNAFinancial Networks: II. Fundamentals of Network Theory and FNA
Financial Networks: II. Fundamentals of Network Theory and FNA
 
A Survey On Secure Cooperative Bait Detection Approach For...
A Survey On Secure Cooperative Bait Detection Approach For...A Survey On Secure Cooperative Bait Detection Approach For...
A Survey On Secure Cooperative Bait Detection Approach For...
 
Jamming aware traffic allocation for multiple-path routing using portfolio se...
Jamming aware traffic allocation for multiple-path routing using portfolio se...Jamming aware traffic allocation for multiple-path routing using portfolio se...
Jamming aware traffic allocation for multiple-path routing using portfolio se...
 
Distributed vertex cover
Distributed vertex coverDistributed vertex cover
Distributed vertex cover
 
Applications of machine learning in Wireless sensor networks.
Applications of machine learning in Wireless sensor networks.Applications of machine learning in Wireless sensor networks.
Applications of machine learning in Wireless sensor networks.
 
A Tale of Data Pattern Discovery in Parallel
A Tale of Data Pattern Discovery in ParallelA Tale of Data Pattern Discovery in Parallel
A Tale of Data Pattern Discovery in Parallel
 
XL-MINER:Prediction
XL-MINER:PredictionXL-MINER:Prediction
XL-MINER:Prediction
 
XL-MINER:Prediction
XL-MINER:PredictionXL-MINER:Prediction
XL-MINER:Prediction
 
Optimizing the Data Collection in Wireless Sensor Network
Optimizing the Data Collection in Wireless Sensor NetworkOptimizing the Data Collection in Wireless Sensor Network
Optimizing the Data Collection in Wireless Sensor Network
 
Deepwalk vs Node2vec
Deepwalk vs Node2vecDeepwalk vs Node2vec
Deepwalk vs Node2vec
 
Description Of A Graph
Description Of A GraphDescription Of A Graph
Description Of A Graph
 
Yarn spark next_gen_hadoop_8_jan_2014
Yarn spark next_gen_hadoop_8_jan_2014Yarn spark next_gen_hadoop_8_jan_2014
Yarn spark next_gen_hadoop_8_jan_2014
 
Spine net learning scale permuted backbone for recognition and localization
Spine net learning scale permuted backbone for recognition and localizationSpine net learning scale permuted backbone for recognition and localization
Spine net learning scale permuted backbone for recognition and localization
 
DDBMS
DDBMSDDBMS
DDBMS
 
SVD and the Netflix Dataset
SVD and the Netflix DatasetSVD and the Netflix Dataset
SVD and the Netflix Dataset
 
P229 godfrey
P229 godfreyP229 godfrey
P229 godfrey
 
PhD Consortium ADBIS presetation.
PhD Consortium ADBIS presetation.PhD Consortium ADBIS presetation.
PhD Consortium ADBIS presetation.
 
MSCX2023_Sergio Gomez_PartI
MSCX2023_Sergio Gomez_PartIMSCX2023_Sergio Gomez_PartI
MSCX2023_Sergio Gomez_PartI
 

More from Kimmo Soramaki

More from Kimmo Soramaki (20)

Applications of Network Theory in Finance
Applications of Network Theory in FinanceApplications of Network Theory in Finance
Applications of Network Theory in Finance
 
Applications of Network Theory in Finance and Production
Applications of Network Theory in Finance and ProductionApplications of Network Theory in Finance and Production
Applications of Network Theory in Finance and Production
 
Global Network of Payment Flows - Presentation at Commerzbank Cash Forum
Global Network of Payment Flows - Presentation at Commerzbank Cash ForumGlobal Network of Payment Flows - Presentation at Commerzbank Cash Forum
Global Network of Payment Flows - Presentation at Commerzbank Cash Forum
 
Visualizing Financial Stress - Talk at European Central Bank
Visualizing Financial Stress - Talk at European Central BankVisualizing Financial Stress - Talk at European Central Bank
Visualizing Financial Stress - Talk at European Central Bank
 
Financial Cartography
Financial CartographyFinancial Cartography
Financial Cartography
 
Financial Cartography at Bogazici University
Financial Cartography at Bogazici UniversityFinancial Cartography at Bogazici University
Financial Cartography at Bogazici University
 
Network Simulations for Business Continuity
Network Simulations for Business ContinuityNetwork Simulations for Business Continuity
Network Simulations for Business Continuity
 
Financial Cartography for Payments and Markets
Financial Cartography for Payments and MarketsFinancial Cartography for Payments and Markets
Financial Cartography for Payments and Markets
 
Quantitative Oversight of Financial Market Infrastructures
Quantitative Oversight of Financial Market InfrastructuresQuantitative Oversight of Financial Market Infrastructures
Quantitative Oversight of Financial Market Infrastructures
 
Emerging Stress Scenarios
Emerging Stress ScenariosEmerging Stress Scenarios
Emerging Stress Scenarios
 
Network Approaches for Interbank Markets
Network Approaches for Interbank MarketsNetwork Approaches for Interbank Markets
Network Approaches for Interbank Markets
 
System shock analysis and complex network effects
System shock analysis and complex network effectsSystem shock analysis and complex network effects
System shock analysis and complex network effects
 
Adaptive Stress Testing
Adaptive Stress TestingAdaptive Stress Testing
Adaptive Stress Testing
 
Illuminating Interconnectedness and Contagion
Illuminating Interconnectedness and ContagionIlluminating Interconnectedness and Contagion
Illuminating Interconnectedness and Contagion
 
Financial Networks and Cartography
Financial Networks and CartographyFinancial Networks and Cartography
Financial Networks and Cartography
 
Financial Networks V - Inferring Links
Financial Networks V - Inferring LinksFinancial Networks V - Inferring Links
Financial Networks V - Inferring Links
 
Financial Cartography - PRMIA Webinar
Financial Cartography - PRMIA WebinarFinancial Cartography - PRMIA Webinar
Financial Cartography - PRMIA Webinar
 
Financial Networks IV. Analyzing and Visualizing Exposures
Financial Networks IV. Analyzing and Visualizing ExposuresFinancial Networks IV. Analyzing and Visualizing Exposures
Financial Networks IV. Analyzing and Visualizing Exposures
 
Financial Networks III. Centrality and Systemic Importance
Financial Networks III. Centrality and Systemic ImportanceFinancial Networks III. Centrality and Systemic Importance
Financial Networks III. Centrality and Systemic Importance
 
Financial Cartography - Center for Financial Research
Financial Cartography - Center for Financial ResearchFinancial Cartography - Center for Financial Research
Financial Cartography - Center for Financial Research
 

Recently uploaded

一比一原版UO毕业证渥太华大学毕业证成绩单如何办理
一比一原版UO毕业证渥太华大学毕业证成绩单如何办理一比一原版UO毕业证渥太华大学毕业证成绩单如何办理
一比一原版UO毕业证渥太华大学毕业证成绩单如何办理
yonemuk
 
NO1 Uk Divorce problem uk all amil baba in karachi,lahore,pakistan talaq ka m...
NO1 Uk Divorce problem uk all amil baba in karachi,lahore,pakistan talaq ka m...NO1 Uk Divorce problem uk all amil baba in karachi,lahore,pakistan talaq ka m...
NO1 Uk Divorce problem uk all amil baba in karachi,lahore,pakistan talaq ka m...
Amil Baba Dawood bangali
 
一比一原版BCU毕业证伯明翰城市大学毕业证成绩单如何办理
一比一原版BCU毕业证伯明翰城市大学毕业证成绩单如何办理一比一原版BCU毕业证伯明翰城市大学毕业证成绩单如何办理
一比一原版BCU毕业证伯明翰城市大学毕业证成绩单如何办理
ydubwyt
 

Recently uploaded (20)

Next Cryptocurrencies to Explode in 2024.pdf
Next Cryptocurrencies to Explode in 2024.pdfNext Cryptocurrencies to Explode in 2024.pdf
Next Cryptocurrencies to Explode in 2024.pdf
 
how to sell pi coins at high rate quickly.
how to sell pi coins at high rate quickly.how to sell pi coins at high rate quickly.
how to sell pi coins at high rate quickly.
 
Can a Pi network coin ever be sold out: I am ready to sell mine.
Can a Pi network coin ever be sold out: I am ready to sell mine.Can a Pi network coin ever be sold out: I am ready to sell mine.
Can a Pi network coin ever be sold out: I am ready to sell mine.
 
Economics and Economic reasoning Chap. 1
Economics and Economic reasoning Chap. 1Economics and Economic reasoning Chap. 1
Economics and Economic reasoning Chap. 1
 
Bitcoin Masterclass TechweekNZ v3.1.pptx
Bitcoin Masterclass TechweekNZ v3.1.pptxBitcoin Masterclass TechweekNZ v3.1.pptx
Bitcoin Masterclass TechweekNZ v3.1.pptx
 
The new type of smart, sustainable entrepreneurship and the next day | Europe...
The new type of smart, sustainable entrepreneurship and the next day | Europe...The new type of smart, sustainable entrepreneurship and the next day | Europe...
The new type of smart, sustainable entrepreneurship and the next day | Europe...
 
how can i make money selling pi coins in 2024
how can i make money selling pi coins in 2024how can i make money selling pi coins in 2024
how can i make money selling pi coins in 2024
 
how can I transfer pi coins to someone in a different country.
how can I transfer pi coins to someone in a different country.how can I transfer pi coins to someone in a different country.
how can I transfer pi coins to someone in a different country.
 
Introduction to Economics II Chapter 25 Production and Growth.pdf
Introduction to Economics II Chapter 25 Production and Growth.pdfIntroduction to Economics II Chapter 25 Production and Growth.pdf
Introduction to Economics II Chapter 25 Production and Growth.pdf
 
一比一原版UO毕业证渥太华大学毕业证成绩单如何办理
一比一原版UO毕业证渥太华大学毕业证成绩单如何办理一比一原版UO毕业证渥太华大学毕业证成绩单如何办理
一比一原版UO毕业证渥太华大学毕业证成绩单如何办理
 
NO1 Uk Divorce problem uk all amil baba in karachi,lahore,pakistan talaq ka m...
NO1 Uk Divorce problem uk all amil baba in karachi,lahore,pakistan talaq ka m...NO1 Uk Divorce problem uk all amil baba in karachi,lahore,pakistan talaq ka m...
NO1 Uk Divorce problem uk all amil baba in karachi,lahore,pakistan talaq ka m...
 
Commercial Bank Economic Capsule - May 2024
Commercial Bank Economic Capsule - May 2024Commercial Bank Economic Capsule - May 2024
Commercial Bank Economic Capsule - May 2024
 
how can i trade pi coins for Bitcoin easily.
how can i trade pi coins for Bitcoin easily.how can i trade pi coins for Bitcoin easily.
how can i trade pi coins for Bitcoin easily.
 
Monthly Market Risk Update: May 2024 [SlideShare]
Monthly Market Risk Update: May 2024 [SlideShare]Monthly Market Risk Update: May 2024 [SlideShare]
Monthly Market Risk Update: May 2024 [SlideShare]
 
Indirect tax .pptx Supply under GST, Charges of GST
Indirect tax .pptx  Supply under GST, Charges of GSTIndirect tax .pptx  Supply under GST, Charges of GST
Indirect tax .pptx Supply under GST, Charges of GST
 
How can I sell my pi coins in Indonesia?
How can I  sell my pi coins in Indonesia?How can I  sell my pi coins in Indonesia?
How can I sell my pi coins in Indonesia?
 
when officially can i withdraw my pi Network coins.
when officially can i withdraw my pi Network coins.when officially can i withdraw my pi Network coins.
when officially can i withdraw my pi Network coins.
 
is it possible to sell pi network coin in 2024.
is it possible to sell pi network coin in 2024.is it possible to sell pi network coin in 2024.
is it possible to sell pi network coin in 2024.
 
一比一原版BCU毕业证伯明翰城市大学毕业证成绩单如何办理
一比一原版BCU毕业证伯明翰城市大学毕业证成绩单如何办理一比一原版BCU毕业证伯明翰城市大学毕业证成绩单如何办理
一比一原版BCU毕业证伯明翰城市大学毕业证成绩单如何办理
 
Summary of financial results for 1Q2024
Summary of financial  results for 1Q2024Summary of financial  results for 1Q2024
Summary of financial results for 1Q2024
 

Financial Networks VI - Correlation Networks

  • 1. Center for Financial Studies at the Goethe University PhD Mini-course Frankfurt, 25 January 2013 Financial Networks VI. Correlation Networks Dr. Kimmo Soramäki Founder and CEO FNA, www.fna.fi
  • 2. Agenda V. Inferring Links • Prices and returns • Controlling for common factors • Correlation and dependence • Significant correlations • Multiple Comparisons VI. Correlation Networks • Distance and Hierarchical Clustering • Minimum Spanning Tree & PMFG • Other filtering • Layout algorithms 2
  • 3. Hierarchical structure in financial markets 3
  • 4. Minimum Spanning Tree A spanning tree of a graph is a subgraph that: 1. is a tree and 2. connects all the nodes together Length of a tree is the sum of its links. Minimum spanning tree (MST) is a spanning tree with shortest length. MST reflects the hierarchical structure of the correlation matrix
  • 5. MST and Hierarchical Structure Source: R.N. Mantegna (1999). Hierarchical structure in nancial markets, Eur. Phys. J. B 11, 193-197 5
  • 6. 36 Single Linkage Clustering • A method for hierarchical clustering • Clusters based on similarity or distance • SLINK algorithm R. Sibson (1973). SLINK: an optimally efficient algorithm for the single-link cluster method. The Computer Journal (British Computer Society) 16 (1): 30–34. 6
  • 7. Example # build network from correlations buildbycorrelationd -file daxreturns-2011-recon.csv -missing Alert -preserve false # calculate distance corrdistance -p correlation -method gower # calculate single linkage clistering slink -p corrdistance # create heatmaps heatmap -sortv vertex_id -p correlation -symmetric true -cellsizedefault 13 - transition 0 -cellhover correlation -palette darkblue-lightgray-darkred - colordomain (-1)-1 -saveas daxheat-slink-Y 7
  • 8. Unordered, Principal Ordered by Cluster, Principal Component Removed Component Removed 8
  • 9. Radial tree -layout • Calculates coordinates for radial layout as presented in Bachmaier, Brandes and Schlieper (2005) • The layout allows definition of each arc length • Specific parameters of command radialtreeviz: – Arc length property (-p) : Arc property defining arc length. Optional. – Root vertex (-rootvertex) : Id of root vertex. The root vertex is placed in the middle of the screen. Due to the repositioning of the tree, nodes may be placed outside the canvas in other than the first network. Optional. – Optimal rotation (-rotation) : Rotates layout to minimize sum of vertex distances between subsequent networks. Optional. By default 'false'. – Scaling (-scale) : Scale of visualization: value/pixel. Christian Bachmaier, Ulrik Brandes, and Barbara Schlieper (2005). Drawing Phylogenetic Trees. Department of Computer & Information Science, University of 9 Konstanz, Germany
  • 10. Putting it all together # build network from correlations buildbycorrelationd -file daxreturns-2011.csv -missing Alert -savestdev -savereturns -preserve false # calculate distance corrdistance -p correlation -method gower # calculate single linkage clistering minst -p corrdistance # drop arcs not in MST dropa -e minst=false # calculate absolute correlation calcap -e 1-abs(correlation) -saveas vizdistance # create heatmaps radialtreeviz -p vizdistance -vlabel vertex_id -vsize stdev -transition 3000 -ahover correlation -saveas daxviz-MST 10
  • 11. Asset Trees Size of node reflects volatility (variance) of returns Links between nodes reflect 'backbone' correlations - short link = high correlation - long link = low correlation 11
  • 12. Circle Tree -visualization • Calculates coordinates for circle tree layout as presented in Bachmaier, Brandes and Schlieper (2005) • As before but instead of radialtreeviz: circletreeviz -vlabel vertex_id -vsize stdev -transition 3000 -ahover correlation -saveas daxviz-MST-circle 12
  • 13. Planar Maximally Filtered Graph Node size scales with degree • A complex graph with loops and cliques of up to 4 elements. It can be drawn on a planar surface without link crossings. • MST is contained in PMFG M. Tumminello, T. Ast, T. Di Matteo and R. N. Mantegna (2005). A Tool for Filtering Information in Complex Systems. PNAS vol. 102 no. 30 pp. 10421–10426 13
  • 14. PMFG -command # build network from correlations buildbycorrelationd -file daxreturns-2011.csv -missing Alert -savestdev -savereturns -preserve false # calculate distance corrdistance -p correlation -method gower # calculate single linkage clistering pmfg -p corrdistance # drop arcs not in MST dropa -e pmfg=false # calculate 1-absolute correlation calcap -e abs(correlation) -saveas vizdistance # calculate degree degree # create heatmaps frviz -vlabel vertex_id -vsize stdev -atransparency vizdistance -ahover correlation -transition 3000 -ahover correlation - arrows false -saveas daxviz-PMFG 14
  • 15. Partial Correlation • Measures the degree of association between two random variables • What is the direct relationship between Adidas and Allianz, controlling for BASF, BAYER, ... ? • We build regression models for Adidas and Allianz and look at the correlation of their model residuals (i.e. wgat left unexplained by the other factors) -> Partial correltation 15
  • 16. Example # build network from correlations buildbypartialcorrelationd -file daxreturns-2011.csv -missing Alert - savestdev -preserve false # show as heatmap heatmap -sortv vertex_id -p partial_correlation -symmetric true - cellsizedefault 13 -transition 0 -cellhover partial_correlation -palette darkblue-lightgray-darkred -colordomain (-1)-1 -saveas daxheat- partial-Y 16
  • 17. Correlations Partial Correlations 17
  • 18. NETS • Network Estimation for Time- Series • Forthcoming paper by Barigozzi and Brownlees • Estimates an unknown network structure from multivariate data • Captures both comtemporenous and serial dependence (partial correlations and lead/lag effects) 18
  • 19. Correlation filtering PMFG Balance between too much and too little information One of many methods to create networks from correlation/distance matrices – PMFGs, Partial Correlation Networks, Influence Networks, Granger Influence Network Causality, NETS, etc. New graph, information-theory, economics & statistics -based models are being actively developed 19
  • 20. Sammon’s Projection Proposed by John W. Sammon in IEEE Transactions on Computers 18: 401–409 (1969) A nonlinear projection method to map a high dimensional space onto a space of lower dimensionality. Example: Iris Setosa Iris Versicolor Iris Virginica 20
  • 21. Example # build network from correlations buildbycorrelationd -file daxreturns-2011.csv -missing Alert -savestdev -savereturns - preserve false # calculate distance corrdistance -p correlation -method gower # Calculate sammonlayout sammonlayouta -p corrdistance -saveerror true # Sum up error sumaforv -p error -saveas error # create heatmaps sammonaviz -p corrdistance -vlabel vertex_id -vsize error -transition 3000 -ahover error -saveas daxviz-Sammon-Y 21
  • 23. Tutorials • Tutorial 1 – Loading Networks into FNA • Tutorial 2 – Managing Data in FNA • Tutorial 3 – Network Summary Measures • Tutorial 4 – Centrality Measures • Tutorial 5 – Connectedness and Components • Tutorial 6 – Network Visualization • Tutorial 7 – Correlation Networks • Tutorial 8 – Payment System Simulations • Tutorial 9 – Analyzing Cross-Border Banking Exposures 23
  • 24. Blog, Library and Demos at www.fna.fi Dr. Kimmo Soramäki kimmo@soramaki.net Twitter: soramaki