A Bayesian View of Challenges in Feature Selection: Feature Aggregation, Multiple Targets, Redundancy and Interaction

Peter Antal, Andras Millinghoffer, Gábor Hullám, Csaba Szalai, András Falus
Proceedings of the Workshop on New Challenges for Feature Selection in Data Mining and Knowledge Discovery at ECML/PKDD 2008, PMLR 4:74-89, 2008.

Abstract

In the paper we discuss applications of the Bayesian approach to new challenges in relevance analysis. Earlier, we formulated a Bayesian approach to Feature Subset Selection using Bayesian networks to jointly estimate the posteriors of Markov Blanket Memberships (MBMs), Markov Blanket Sets (MBSs), and Markov Blanket Graphs (MBGs) for a given target variable. These results of the Bayesian Multilevel Analysis of relevance (BMLA) correspond respectively to a model-based pairwise relevance, relevance of sets, and to the interaction models of relevant variables. Now we formulate refined levels in BMLA by introducing the concepts of k-MBSs and k-MBGs, which are intermediate, scalable model properties expressing relevance. We consider the extension of BMLA to multiple targets. We introduce and investigate a score for feature redundancy and interaction based on the decomposability of the structure posterior. Finally, we overview the problems of conditional and contextual relevance. We demonstrate the use of concepts and methods in the field of genomics of asthma.

Cite this Paper

BibTeX
@InProceedings{pmlr-v4-antal08a, title = {A Bayesian View of Challenges in Feature Selection: Feature Aggregation, Multiple Targets, Redundancy and Interaction}, author = {Antal, Peter and Millinghoffer, Andras and Hullám, Gábor and Szalai, Csaba and Falus, András}, booktitle = {Proceedings of the Workshop on New Challenges for Feature Selection in Data Mining and Knowledge Discovery at ECML/PKDD 2008}, pages = {74--89}, year = {2008}, editor = {Saeys, Yvan and Liu, Huan and Inza, Iñaki and Wehenkel, Louis and Pee, Yves Van de}, volume = {4}, series = {Proceedings of Machine Learning Research}, address = {Antwerp, Belgium}, month = {15 Sep}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v4/antal08a/antal08a.pdf}, url = {https://proceedings.mlr.press/v4/antal08a.html}, abstract = {In the paper we discuss applications of the Bayesian approach to new challenges in relevance analysis. Earlier, we formulated a Bayesian approach to Feature Subset Selection using Bayesian networks to jointly estimate the posteriors of Markov Blanket Memberships (MBMs), Markov Blanket Sets (MBSs), and Markov Blanket Graphs (MBGs) for a given target variable. These results of the Bayesian Multilevel Analysis of relevance (BMLA) correspond respectively to a model-based pairwise relevance, relevance of sets, and to the interaction models of relevant variables. Now we formulate refined levels in BMLA by introducing the concepts of k-MBSs and k-MBGs, which are intermediate, scalable model properties expressing relevance. We consider the extension of BMLA to multiple targets. We introduce and investigate a score for feature redundancy and interaction based on the decomposability of the structure posterior. Finally, we overview the problems of conditional and contextual relevance. We demonstrate the use of concepts and methods in the field of genomics of asthma.} }
Endnote
%0 Conference Paper %T A Bayesian View of Challenges in Feature Selection: Feature Aggregation, Multiple Targets, Redundancy and Interaction %A Peter Antal %A Andras Millinghoffer %A Gábor Hullám %A Csaba Szalai %A András Falus %B Proceedings of the Workshop on New Challenges for Feature Selection in Data Mining and Knowledge Discovery at ECML/PKDD 2008 %C Proceedings of Machine Learning Research %D 2008 %E Yvan Saeys %E Huan Liu %E Iñaki Inza %E Louis Wehenkel %E Yves Van de Pee %F pmlr-v4-antal08a %I PMLR %P 74--89 %U https://proceedings.mlr.press/v4/antal08a.html %V 4 %X In the paper we discuss applications of the Bayesian approach to new challenges in relevance analysis. Earlier, we formulated a Bayesian approach to Feature Subset Selection using Bayesian networks to jointly estimate the posteriors of Markov Blanket Memberships (MBMs), Markov Blanket Sets (MBSs), and Markov Blanket Graphs (MBGs) for a given target variable. These results of the Bayesian Multilevel Analysis of relevance (BMLA) correspond respectively to a model-based pairwise relevance, relevance of sets, and to the interaction models of relevant variables. Now we formulate refined levels in BMLA by introducing the concepts of k-MBSs and k-MBGs, which are intermediate, scalable model properties expressing relevance. We consider the extension of BMLA to multiple targets. We introduce and investigate a score for feature redundancy and interaction based on the decomposability of the structure posterior. Finally, we overview the problems of conditional and contextual relevance. We demonstrate the use of concepts and methods in the field of genomics of asthma.
RIS
TY - CPAPER TI - A Bayesian View of Challenges in Feature Selection: Feature Aggregation, Multiple Targets, Redundancy and Interaction AU - Peter Antal AU - Andras Millinghoffer AU - Gábor Hullám AU - Csaba Szalai AU - András Falus BT - Proceedings of the Workshop on New Challenges for Feature Selection in Data Mining and Knowledge Discovery at ECML/PKDD 2008 DA - 2008/09/11 ED - Yvan Saeys ED - Huan Liu ED - Iñaki Inza ED - Louis Wehenkel ED - Yves Van de Pee ID - pmlr-v4-antal08a PB - PMLR DP - Proceedings of Machine Learning Research VL - 4 SP - 74 EP - 89 L1 - http://proceedings.mlr.press/v4/antal08a/antal08a.pdf UR - https://proceedings.mlr.press/v4/antal08a.html AB - In the paper we discuss applications of the Bayesian approach to new challenges in relevance analysis. Earlier, we formulated a Bayesian approach to Feature Subset Selection using Bayesian networks to jointly estimate the posteriors of Markov Blanket Memberships (MBMs), Markov Blanket Sets (MBSs), and Markov Blanket Graphs (MBGs) for a given target variable. These results of the Bayesian Multilevel Analysis of relevance (BMLA) correspond respectively to a model-based pairwise relevance, relevance of sets, and to the interaction models of relevant variables. Now we formulate refined levels in BMLA by introducing the concepts of k-MBSs and k-MBGs, which are intermediate, scalable model properties expressing relevance. We consider the extension of BMLA to multiple targets. We introduce and investigate a score for feature redundancy and interaction based on the decomposability of the structure posterior. Finally, we overview the problems of conditional and contextual relevance. We demonstrate the use of concepts and methods in the field of genomics of asthma. ER -
APA
Antal, P., Millinghoffer, A., Hullám, G., Szalai, C. & Falus, A.. (2008). A Bayesian View of Challenges in Feature Selection: Feature Aggregation, Multiple Targets, Redundancy and Interaction. Proceedings of the Workshop on New Challenges for Feature Selection in Data Mining and Knowledge Discovery at ECML/PKDD 2008, in Proceedings of Machine Learning Research 4:74-89 Available from https://proceedings.mlr.press/v4/antal08a.html.

Related Material