The Inverse Regression Topic Model

Maxim Rabinovich, David Blei
Proceedings of the 31st International Conference on Machine Learning, PMLR 32(1):199-207, 2014.

Abstract

\citettaddy13mnir proposed multinomial inverse regression (MNIR) as a new model of annotated text based on the influence of metadata and response variables on the distribution of words in a document. While effective, MNIR has no way to exploit structure in the corpus to improve its predictions or facilitate exploratory data analysis. On the other hand, traditional probabilistic topic models (like latent Dirichlet allocation) capture natural heterogeneity in a collection but do not account for external variables. In this paper, we introduce the inverse regression topic model (IRTM), a mixed-membership extension of MNIR that combines the strengths of both methodologies. We present two inference algorithms for the IRTM: an efficient batch estimation algorithm and an online variant, which is suitable for large corpora. We apply these methods to a corpus of 73K Congressional press releases and another of 150K Yelp reviews, demonstrating that the IRTM outperforms both MNIR and supervised topic models on the prediction task. Further, we give examples showing that the IRTM enables systematic discovery of in-topic lexical variation, which is not possible with previous supervised topic models.

Cite this Paper


BibTeX
@InProceedings{pmlr-v32-rabinovich14, title = {The Inverse Regression Topic Model}, author = {Rabinovich, Maxim and Blei, David}, booktitle = {Proceedings of the 31st International Conference on Machine Learning}, pages = {199--207}, year = {2014}, editor = {Xing, Eric P. and Jebara, Tony}, volume = {32}, number = {1}, series = {Proceedings of Machine Learning Research}, address = {Bejing, China}, month = {22--24 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v32/rabinovich14.pdf}, url = {https://proceedings.mlr.press/v32/rabinovich14.html}, abstract = {\citettaddy13mnir proposed multinomial inverse regression (MNIR) as a new model of annotated text based on the influence of metadata and response variables on the distribution of words in a document. While effective, MNIR has no way to exploit structure in the corpus to improve its predictions or facilitate exploratory data analysis. On the other hand, traditional probabilistic topic models (like latent Dirichlet allocation) capture natural heterogeneity in a collection but do not account for external variables. In this paper, we introduce the inverse regression topic model (IRTM), a mixed-membership extension of MNIR that combines the strengths of both methodologies. We present two inference algorithms for the IRTM: an efficient batch estimation algorithm and an online variant, which is suitable for large corpora. We apply these methods to a corpus of 73K Congressional press releases and another of 150K Yelp reviews, demonstrating that the IRTM outperforms both MNIR and supervised topic models on the prediction task. Further, we give examples showing that the IRTM enables systematic discovery of in-topic lexical variation, which is not possible with previous supervised topic models.} }
Endnote
%0 Conference Paper %T The Inverse Regression Topic Model %A Maxim Rabinovich %A David Blei %B Proceedings of the 31st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2014 %E Eric P. Xing %E Tony Jebara %F pmlr-v32-rabinovich14 %I PMLR %P 199--207 %U https://proceedings.mlr.press/v32/rabinovich14.html %V 32 %N 1 %X \citettaddy13mnir proposed multinomial inverse regression (MNIR) as a new model of annotated text based on the influence of metadata and response variables on the distribution of words in a document. While effective, MNIR has no way to exploit structure in the corpus to improve its predictions or facilitate exploratory data analysis. On the other hand, traditional probabilistic topic models (like latent Dirichlet allocation) capture natural heterogeneity in a collection but do not account for external variables. In this paper, we introduce the inverse regression topic model (IRTM), a mixed-membership extension of MNIR that combines the strengths of both methodologies. We present two inference algorithms for the IRTM: an efficient batch estimation algorithm and an online variant, which is suitable for large corpora. We apply these methods to a corpus of 73K Congressional press releases and another of 150K Yelp reviews, demonstrating that the IRTM outperforms both MNIR and supervised topic models on the prediction task. Further, we give examples showing that the IRTM enables systematic discovery of in-topic lexical variation, which is not possible with previous supervised topic models.
RIS
TY - CPAPER TI - The Inverse Regression Topic Model AU - Maxim Rabinovich AU - David Blei BT - Proceedings of the 31st International Conference on Machine Learning DA - 2014/01/27 ED - Eric P. Xing ED - Tony Jebara ID - pmlr-v32-rabinovich14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 32 IS - 1 SP - 199 EP - 207 L1 - http://proceedings.mlr.press/v32/rabinovich14.pdf UR - https://proceedings.mlr.press/v32/rabinovich14.html AB - \citettaddy13mnir proposed multinomial inverse regression (MNIR) as a new model of annotated text based on the influence of metadata and response variables on the distribution of words in a document. While effective, MNIR has no way to exploit structure in the corpus to improve its predictions or facilitate exploratory data analysis. On the other hand, traditional probabilistic topic models (like latent Dirichlet allocation) capture natural heterogeneity in a collection but do not account for external variables. In this paper, we introduce the inverse regression topic model (IRTM), a mixed-membership extension of MNIR that combines the strengths of both methodologies. We present two inference algorithms for the IRTM: an efficient batch estimation algorithm and an online variant, which is suitable for large corpora. We apply these methods to a corpus of 73K Congressional press releases and another of 150K Yelp reviews, demonstrating that the IRTM outperforms both MNIR and supervised topic models on the prediction task. Further, we give examples showing that the IRTM enables systematic discovery of in-topic lexical variation, which is not possible with previous supervised topic models. ER -
APA
Rabinovich, M. & Blei, D.. (2014). The Inverse Regression Topic Model. Proceedings of the 31st International Conference on Machine Learning, in Proceedings of Machine Learning Research 32(1):199-207 Available from https://proceedings.mlr.press/v32/rabinovich14.html.

Related Material