Subset Infinite Relational Models

Katsuhiko Ishiguro, Naonori Ueda, Hiroshi Sawada
; Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, PMLR 22:547-555, 2012.

Abstract

We propose a new probabilistic generative model for analyzing sparse and noisy pairwise relational data, such as friend-links on SNSs and customer records in online shops. Real-world relational data often include a large portion of non-informative pairwise data entries. Many existing stochastic blockmodels suffer from these irrelevant data entries because of their rather simpler forms of priors. The proposed model newly incorporates a latent variable that explicitly indicates whether each data entry is relevant or not to diminish the bad effects associated with such irrelevant data. Through experimental results using synthetic and real data sets, we show that the proposed model can extract clusters with stronger relations among data within the cluster than clusters obtained by the conventional model.

Cite this Paper


BibTeX
@InProceedings{pmlr-v22-ishiguro12, title = {Subset Infinite Relational Models}, author = {Katsuhiko Ishiguro and Naonori Ueda and Hiroshi Sawada}, booktitle = {Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics}, pages = {547--555}, year = {2012}, editor = {Neil D. Lawrence and Mark Girolami}, volume = {22}, series = {Proceedings of Machine Learning Research}, address = {La Palma, Canary Islands}, month = {21--23 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v22/ishiguro12/ishiguro12.pdf}, url = {http://proceedings.mlr.press/v22/ishiguro12.html}, abstract = {We propose a new probabilistic generative model for analyzing sparse and noisy pairwise relational data, such as friend-links on SNSs and customer records in online shops. Real-world relational data often include a large portion of non-informative pairwise data entries. Many existing stochastic blockmodels suffer from these irrelevant data entries because of their rather simpler forms of priors. The proposed model newly incorporates a latent variable that explicitly indicates whether each data entry is relevant or not to diminish the bad effects associated with such irrelevant data. Through experimental results using synthetic and real data sets, we show that the proposed model can extract clusters with stronger relations among data within the cluster than clusters obtained by the conventional model.} }
Endnote
%0 Conference Paper %T Subset Infinite Relational Models %A Katsuhiko Ishiguro %A Naonori Ueda %A Hiroshi Sawada %B Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2012 %E Neil D. Lawrence %E Mark Girolami %F pmlr-v22-ishiguro12 %I PMLR %J Proceedings of Machine Learning Research %P 547--555 %U http://proceedings.mlr.press %V 22 %W PMLR %X We propose a new probabilistic generative model for analyzing sparse and noisy pairwise relational data, such as friend-links on SNSs and customer records in online shops. Real-world relational data often include a large portion of non-informative pairwise data entries. Many existing stochastic blockmodels suffer from these irrelevant data entries because of their rather simpler forms of priors. The proposed model newly incorporates a latent variable that explicitly indicates whether each data entry is relevant or not to diminish the bad effects associated with such irrelevant data. Through experimental results using synthetic and real data sets, we show that the proposed model can extract clusters with stronger relations among data within the cluster than clusters obtained by the conventional model.
RIS
TY - CPAPER TI - Subset Infinite Relational Models AU - Katsuhiko Ishiguro AU - Naonori Ueda AU - Hiroshi Sawada BT - Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics PY - 2012/03/21 DA - 2012/03/21 ED - Neil D. Lawrence ED - Mark Girolami ID - pmlr-v22-ishiguro12 PB - PMLR SP - 547 DP - PMLR EP - 555 L1 - http://proceedings.mlr.press/v22/ishiguro12/ishiguro12.pdf UR - http://proceedings.mlr.press/v22/ishiguro12.html AB - We propose a new probabilistic generative model for analyzing sparse and noisy pairwise relational data, such as friend-links on SNSs and customer records in online shops. Real-world relational data often include a large portion of non-informative pairwise data entries. Many existing stochastic blockmodels suffer from these irrelevant data entries because of their rather simpler forms of priors. The proposed model newly incorporates a latent variable that explicitly indicates whether each data entry is relevant or not to diminish the bad effects associated with such irrelevant data. Through experimental results using synthetic and real data sets, we show that the proposed model can extract clusters with stronger relations among data within the cluster than clusters obtained by the conventional model. ER -
APA
Ishiguro, K., Ueda, N. & Sawada, H.. (2012). Subset Infinite Relational Models. Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, in PMLR 22:547-555

Related Material