SlideShare a Scribd company logo
1 of 1
Download to read offline
ECWAY TECHNOLOGIES
IEEE PROJECTS & SOFTWARE DEVELOPMENTS
OUR OFFICES @ CHENNAI / TRICHY / KARUR / ERODE / MADURAI / SALEM / COIMBATORE
CELL: +91 98949 17187, +91 875487 2111 / 3111 / 4111 / 5111 / 6111
VISIT: www.ecwayprojects.com MAIL TO: ecwaytechnologies@gmail.com

A GRAPH-BASED CONSENSUS MAXIMIZATION APPROACH FOR COMBINING
MULTIPLE SUPERVISED AND UNSUPERVISED MODELS

ABSTRACT:
Ensemble learning has emerged as a powerful method for combining multiple models. Wellknown methods, such as bagging, boosting, and model averaging, have been shown to improve
accuracy and robustness over single models. However, due to the high costs of manual labeling,
it is hard to obtain sufficient and reliable labeled data for effective training. Meanwhile, lots of
unlabeled data exist in these sources, and we can readily obtain multiple unsupervised models.
Although unsupervised models do not directly generate a class label prediction for each object,
they provide useful constraints on the joint predictions for a set of related objects. Therefore,
incorporating these unsupervised models into the ensemble of supervised models can lead to
better prediction performance.
In this paper, we study ensemble learning with outputs from multiple supervised and
unsupervised models, a topic where little work has been done. We propose to consolidate a
classification solution by maximizing the consensus among both supervised predictions and
unsupervised constraints. We cast this ensemble task as an optimization problem on a bipartite
graph, where the objective function favors the smoothness of the predictions over the graph, but
penalizes the deviations from the initial labeling provided by the supervised models. We solve
this problem through iterative propagation of probability estimates among neighboring nodes and
prove the optimality of the solution. The proposed method can be interpreted as conducting a
constrained embedding in a transformed space, or a ranking on the graph. Experimental results
on different applications with heterogeneous data sources demonstrate the benefits of the
proposed method over existing alternatives.

More Related Content

What's hot (7)

Numerical methods and its applications
Numerical methods and its applicationsNumerical methods and its applications
Numerical methods and its applications
 
real life application in numerical method
real life application in numerical methodreal life application in numerical method
real life application in numerical method
 
Dadm (lys)
Dadm (lys)Dadm (lys)
Dadm (lys)
 
Meta Learning Shared Hierarchies
Meta Learning Shared HierarchiesMeta Learning Shared Hierarchies
Meta Learning Shared Hierarchies
 
Energy Behaviour and Smart Meters
Energy Behaviour and Smart MetersEnergy Behaviour and Smart Meters
Energy Behaviour and Smart Meters
 
Machine learning
Machine learningMachine learning
Machine learning
 
Poster systems
Poster systemsPoster systems
Poster systems
 

Viewers also liked

A current controller design for current source inverter fed ac machine drive ...
A current controller design for current source inverter fed ac machine drive ...A current controller design for current source inverter fed ac machine drive ...
A current controller design for current source inverter fed ac machine drive ...Ecwaytech
 
2013 2014 ieee vlsi titles
2013 2014 ieee vlsi titles2013 2014 ieee vlsi titles
2013 2014 ieee vlsi titlesEcwaytech
 
2014 ieee project ns2 titles
2014 ieee project ns2 titles2014 ieee project ns2 titles
2014 ieee project ns2 titlesEcwaytech
 
A new dc anti islanding technique of electrolytic capacitor-less photovoltaic...
A new dc anti islanding technique of electrolytic capacitor-less photovoltaic...A new dc anti islanding technique of electrolytic capacitor-less photovoltaic...
A new dc anti islanding technique of electrolytic capacitor-less photovoltaic...Ecwaytech
 
2014 ieee powerelectronics
2014 ieee powerelectronics2014 ieee powerelectronics
2014 ieee powerelectronicsEcwaytech
 
CCNA Routing and Switching Lesson 06 - IOS Basics - Eric Vanderburg
CCNA Routing and Switching Lesson 06 - IOS Basics - Eric VanderburgCCNA Routing and Switching Lesson 06 - IOS Basics - Eric Vanderburg
CCNA Routing and Switching Lesson 06 - IOS Basics - Eric VanderburgEric Vanderburg
 
Idees emploi : synthèse des travaux sur l'emploi des seniors
Idees emploi : synthèse des travaux sur l'emploi des seniorsIdees emploi : synthèse des travaux sur l'emploi des seniors
Idees emploi : synthèse des travaux sur l'emploi des seniorsDaniel Pigeon-Angelini
 
A hybrid multiview stereo algorithm for modeling urban scenes
A hybrid multiview stereo algorithm for modeling urban scenesA hybrid multiview stereo algorithm for modeling urban scenes
A hybrid multiview stereo algorithm for modeling urban scenesEcwaytech
 
2014 ieee embedded projects
2014 ieee embedded projects2014 ieee embedded projects
2014 ieee embedded projectsEcwaytech
 
2014 ieee project matlab titles
2014 ieee project matlab titles2014 ieee project matlab titles
2014 ieee project matlab titlesEcwaytech
 
Gaurav Ahuja- Resume
Gaurav Ahuja- ResumeGaurav Ahuja- Resume
Gaurav Ahuja- ResumeGaurav Ahuja
 
Yitro group 1
Yitro   group 1Yitro   group 1
Yitro group 1el9360
 
Bb Collaborate Web Conferencing Guide
Bb Collaborate Web Conferencing GuideBb Collaborate Web Conferencing Guide
Bb Collaborate Web Conferencing GuideKayAdministrator
 
Languagelab 8.4 - Complex Sentences
Languagelab 8.4 - Complex SentencesLanguagelab 8.4 - Complex Sentences
Languagelab 8.4 - Complex SentencesDesignlab Innovation
 
A proxy based approach to continuous location-based spatial queries in mobile...
A proxy based approach to continuous location-based spatial queries in mobile...A proxy based approach to continuous location-based spatial queries in mobile...
A proxy based approach to continuous location-based spatial queries in mobile...Ecwaytech
 
2014 ieee project java titles
2014 ieee project java titles2014 ieee project java titles
2014 ieee project java titlesEcwaytech
 
2014 ieee project dotnet titles
2014 ieee project dotnet titles2014 ieee project dotnet titles
2014 ieee project dotnet titlesEcwaytech
 

Viewers also liked (19)

A current controller design for current source inverter fed ac machine drive ...
A current controller design for current source inverter fed ac machine drive ...A current controller design for current source inverter fed ac machine drive ...
A current controller design for current source inverter fed ac machine drive ...
 
2013 2014 ieee vlsi titles
2013 2014 ieee vlsi titles2013 2014 ieee vlsi titles
2013 2014 ieee vlsi titles
 
2014 ieee project ns2 titles
2014 ieee project ns2 titles2014 ieee project ns2 titles
2014 ieee project ns2 titles
 
A new dc anti islanding technique of electrolytic capacitor-less photovoltaic...
A new dc anti islanding technique of electrolytic capacitor-less photovoltaic...A new dc anti islanding technique of electrolytic capacitor-less photovoltaic...
A new dc anti islanding technique of electrolytic capacitor-less photovoltaic...
 
2014 ieee powerelectronics
2014 ieee powerelectronics2014 ieee powerelectronics
2014 ieee powerelectronics
 
CCNA Routing and Switching Lesson 06 - IOS Basics - Eric Vanderburg
CCNA Routing and Switching Lesson 06 - IOS Basics - Eric VanderburgCCNA Routing and Switching Lesson 06 - IOS Basics - Eric Vanderburg
CCNA Routing and Switching Lesson 06 - IOS Basics - Eric Vanderburg
 
Idees emploi : synthèse des travaux sur l'emploi des seniors
Idees emploi : synthèse des travaux sur l'emploi des seniorsIdees emploi : synthèse des travaux sur l'emploi des seniors
Idees emploi : synthèse des travaux sur l'emploi des seniors
 
A hybrid multiview stereo algorithm for modeling urban scenes
A hybrid multiview stereo algorithm for modeling urban scenesA hybrid multiview stereo algorithm for modeling urban scenes
A hybrid multiview stereo algorithm for modeling urban scenes
 
2014 ieee embedded projects
2014 ieee embedded projects2014 ieee embedded projects
2014 ieee embedded projects
 
2014 ieee project matlab titles
2014 ieee project matlab titles2014 ieee project matlab titles
2014 ieee project matlab titles
 
Cannon Poster
Cannon PosterCannon Poster
Cannon Poster
 
Gaurav Ahuja- Resume
Gaurav Ahuja- ResumeGaurav Ahuja- Resume
Gaurav Ahuja- Resume
 
Yitro group 1
Yitro   group 1Yitro   group 1
Yitro group 1
 
Bb Collaborate Web Conferencing Guide
Bb Collaborate Web Conferencing GuideBb Collaborate Web Conferencing Guide
Bb Collaborate Web Conferencing Guide
 
Languagelab 8.4 - Complex Sentences
Languagelab 8.4 - Complex SentencesLanguagelab 8.4 - Complex Sentences
Languagelab 8.4 - Complex Sentences
 
A proxy based approach to continuous location-based spatial queries in mobile...
A proxy based approach to continuous location-based spatial queries in mobile...A proxy based approach to continuous location-based spatial queries in mobile...
A proxy based approach to continuous location-based spatial queries in mobile...
 
2014 ieee project java titles
2014 ieee project java titles2014 ieee project java titles
2014 ieee project java titles
 
2014 ieee project dotnet titles
2014 ieee project dotnet titles2014 ieee project dotnet titles
2014 ieee project dotnet titles
 
Lebenslauf 2015
Lebenslauf 2015Lebenslauf 2015
Lebenslauf 2015
 

Similar to Graph-based consensus maximization approach

A graph based consensus maximization approach for combining multiple supervis...
A graph based consensus maximization approach for combining multiple supervis...A graph based consensus maximization approach for combining multiple supervis...
A graph based consensus maximization approach for combining multiple supervis...ecway
 
Java a graph-based consensus maximization approach for combining multiple su...
Java  a graph-based consensus maximization approach for combining multiple su...Java  a graph-based consensus maximization approach for combining multiple su...
Java a graph-based consensus maximization approach for combining multiple su...ecwayerode
 
An Ensemble Approach To Improve Homomorphic Encrypted Data Classification Per...
An Ensemble Approach To Improve Homomorphic Encrypted Data Classification Per...An Ensemble Approach To Improve Homomorphic Encrypted Data Classification Per...
An Ensemble Approach To Improve Homomorphic Encrypted Data Classification Per...IJCI JOURNAL
 
An approach for improved students’ performance prediction using homogeneous ...
An approach for improved students’ performance prediction  using homogeneous ...An approach for improved students’ performance prediction  using homogeneous ...
An approach for improved students’ performance prediction using homogeneous ...IJECEIAES
 
IEEE Datamining 2016 Title and Abstract
IEEE  Datamining 2016 Title and AbstractIEEE  Datamining 2016 Title and Abstract
IEEE Datamining 2016 Title and Abstracttsysglobalsolutions
 
IEEE 2014 MATLAB IMAGE PROCESSING PROJECTS Multi label image categorization w...
IEEE 2014 MATLAB IMAGE PROCESSING PROJECTS Multi label image categorization w...IEEE 2014 MATLAB IMAGE PROCESSING PROJECTS Multi label image categorization w...
IEEE 2014 MATLAB IMAGE PROCESSING PROJECTS Multi label image categorization w...IEEEBEBTECHSTUDENTPROJECTS
 
Ieee transactions on 2018 knowledge and data engineering topics with abstract .
Ieee transactions on 2018 knowledge and data engineering topics with abstract .Ieee transactions on 2018 knowledge and data engineering topics with abstract .
Ieee transactions on 2018 knowledge and data engineering topics with abstract .tsysglobalsolutions
 
Strategies for energy efficient resource management of hybrid programming models
Strategies for energy efficient resource management of hybrid programming modelsStrategies for energy efficient resource management of hybrid programming models
Strategies for energy efficient resource management of hybrid programming modelsEcwayt
 
Dotnet strategies for energy-efficient resource management of hybrid program...
Dotnet  strategies for energy-efficient resource management of hybrid program...Dotnet  strategies for energy-efficient resource management of hybrid program...
Dotnet strategies for energy-efficient resource management of hybrid program...Ecwaytech
 
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...IAEME Publication
 
Dotnet maximum likelihood estimation from uncertain data in the belief funct...
Dotnet  maximum likelihood estimation from uncertain data in the belief funct...Dotnet  maximum likelihood estimation from uncertain data in the belief funct...
Dotnet maximum likelihood estimation from uncertain data in the belief funct...Ecwayt
 
Dotnet maximum likelihood estimation from uncertain data in the belief funct...
Dotnet  maximum likelihood estimation from uncertain data in the belief funct...Dotnet  maximum likelihood estimation from uncertain data in the belief funct...
Dotnet maximum likelihood estimation from uncertain data in the belief funct...Ecwaytech
 
Maximum likelihood estimation from uncertain data in the belief function fram...
Maximum likelihood estimation from uncertain data in the belief function fram...Maximum likelihood estimation from uncertain data in the belief function fram...
Maximum likelihood estimation from uncertain data in the belief function fram...ecwayprojects
 
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNINGUNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNINGgerogepatton
 
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNINGUNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNINGijaia
 
When deep learners change their mind learning dynamics for active learning
When deep learners change their mind  learning dynamics for active learningWhen deep learners change their mind  learning dynamics for active learning
When deep learners change their mind learning dynamics for active learningDevansh16
 
IEEE Pattern analysis and machine intelligence 2016 Title and Abstract
IEEE Pattern analysis and machine intelligence 2016 Title and AbstractIEEE Pattern analysis and machine intelligence 2016 Title and Abstract
IEEE Pattern analysis and machine intelligence 2016 Title and Abstracttsysglobalsolutions
 
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKSCAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKSgerogepatton
 
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKSCAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKSijaia
 
Clustering large probabilistic graphs
Clustering large probabilistic graphsClustering large probabilistic graphs
Clustering large probabilistic graphsEcway2004
 

Similar to Graph-based consensus maximization approach (20)

A graph based consensus maximization approach for combining multiple supervis...
A graph based consensus maximization approach for combining multiple supervis...A graph based consensus maximization approach for combining multiple supervis...
A graph based consensus maximization approach for combining multiple supervis...
 
Java a graph-based consensus maximization approach for combining multiple su...
Java  a graph-based consensus maximization approach for combining multiple su...Java  a graph-based consensus maximization approach for combining multiple su...
Java a graph-based consensus maximization approach for combining multiple su...
 
An Ensemble Approach To Improve Homomorphic Encrypted Data Classification Per...
An Ensemble Approach To Improve Homomorphic Encrypted Data Classification Per...An Ensemble Approach To Improve Homomorphic Encrypted Data Classification Per...
An Ensemble Approach To Improve Homomorphic Encrypted Data Classification Per...
 
An approach for improved students’ performance prediction using homogeneous ...
An approach for improved students’ performance prediction  using homogeneous ...An approach for improved students’ performance prediction  using homogeneous ...
An approach for improved students’ performance prediction using homogeneous ...
 
IEEE Datamining 2016 Title and Abstract
IEEE  Datamining 2016 Title and AbstractIEEE  Datamining 2016 Title and Abstract
IEEE Datamining 2016 Title and Abstract
 
IEEE 2014 MATLAB IMAGE PROCESSING PROJECTS Multi label image categorization w...
IEEE 2014 MATLAB IMAGE PROCESSING PROJECTS Multi label image categorization w...IEEE 2014 MATLAB IMAGE PROCESSING PROJECTS Multi label image categorization w...
IEEE 2014 MATLAB IMAGE PROCESSING PROJECTS Multi label image categorization w...
 
Ieee transactions on 2018 knowledge and data engineering topics with abstract .
Ieee transactions on 2018 knowledge and data engineering topics with abstract .Ieee transactions on 2018 knowledge and data engineering topics with abstract .
Ieee transactions on 2018 knowledge and data engineering topics with abstract .
 
Strategies for energy efficient resource management of hybrid programming models
Strategies for energy efficient resource management of hybrid programming modelsStrategies for energy efficient resource management of hybrid programming models
Strategies for energy efficient resource management of hybrid programming models
 
Dotnet strategies for energy-efficient resource management of hybrid program...
Dotnet  strategies for energy-efficient resource management of hybrid program...Dotnet  strategies for energy-efficient resource management of hybrid program...
Dotnet strategies for energy-efficient resource management of hybrid program...
 
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
AN IMPROVED METHOD FOR IDENTIFYING WELL-TEST INTERPRETATION MODEL BASED ON AG...
 
Dotnet maximum likelihood estimation from uncertain data in the belief funct...
Dotnet  maximum likelihood estimation from uncertain data in the belief funct...Dotnet  maximum likelihood estimation from uncertain data in the belief funct...
Dotnet maximum likelihood estimation from uncertain data in the belief funct...
 
Dotnet maximum likelihood estimation from uncertain data in the belief funct...
Dotnet  maximum likelihood estimation from uncertain data in the belief funct...Dotnet  maximum likelihood estimation from uncertain data in the belief funct...
Dotnet maximum likelihood estimation from uncertain data in the belief funct...
 
Maximum likelihood estimation from uncertain data in the belief function fram...
Maximum likelihood estimation from uncertain data in the belief function fram...Maximum likelihood estimation from uncertain data in the belief function fram...
Maximum likelihood estimation from uncertain data in the belief function fram...
 
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNINGUNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
 
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNINGUNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
UNCERTAINTY ESTIMATION IN NEURAL NETWORKS THROUGH MULTI-TASK LEARNING
 
When deep learners change their mind learning dynamics for active learning
When deep learners change their mind  learning dynamics for active learningWhen deep learners change their mind  learning dynamics for active learning
When deep learners change their mind learning dynamics for active learning
 
IEEE Pattern analysis and machine intelligence 2016 Title and Abstract
IEEE Pattern analysis and machine intelligence 2016 Title and AbstractIEEE Pattern analysis and machine intelligence 2016 Title and Abstract
IEEE Pattern analysis and machine intelligence 2016 Title and Abstract
 
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKSCAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
 
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKSCAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
CAUSALITY LEARNING WITH WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKS
 
Clustering large probabilistic graphs
Clustering large probabilistic graphsClustering large probabilistic graphs
Clustering large probabilistic graphs
 

Graph-based consensus maximization approach

  • 1. ECWAY TECHNOLOGIES IEEE PROJECTS & SOFTWARE DEVELOPMENTS OUR OFFICES @ CHENNAI / TRICHY / KARUR / ERODE / MADURAI / SALEM / COIMBATORE CELL: +91 98949 17187, +91 875487 2111 / 3111 / 4111 / 5111 / 6111 VISIT: www.ecwayprojects.com MAIL TO: ecwaytechnologies@gmail.com A GRAPH-BASED CONSENSUS MAXIMIZATION APPROACH FOR COMBINING MULTIPLE SUPERVISED AND UNSUPERVISED MODELS ABSTRACT: Ensemble learning has emerged as a powerful method for combining multiple models. Wellknown methods, such as bagging, boosting, and model averaging, have been shown to improve accuracy and robustness over single models. However, due to the high costs of manual labeling, it is hard to obtain sufficient and reliable labeled data for effective training. Meanwhile, lots of unlabeled data exist in these sources, and we can readily obtain multiple unsupervised models. Although unsupervised models do not directly generate a class label prediction for each object, they provide useful constraints on the joint predictions for a set of related objects. Therefore, incorporating these unsupervised models into the ensemble of supervised models can lead to better prediction performance. In this paper, we study ensemble learning with outputs from multiple supervised and unsupervised models, a topic where little work has been done. We propose to consolidate a classification solution by maximizing the consensus among both supervised predictions and unsupervised constraints. We cast this ensemble task as an optimization problem on a bipartite graph, where the objective function favors the smoothness of the predictions over the graph, but penalizes the deviations from the initial labeling provided by the supervised models. We solve this problem through iterative propagation of probability estimates among neighboring nodes and prove the optimality of the solution. The proposed method can be interpreted as conducting a constrained embedding in a transformed space, or a ranking on the graph. Experimental results on different applications with heterogeneous data sources demonstrate the benefits of the proposed method over existing alternatives.