Relational topic models (RTM) have been widely used to discover hidden topics in a collection of networked documents. In this paper, we introduce the class of Constrained Relational Topic Models (CRTM), a semi-supervised extension of RTM that, apart from modeling the structure of the document network, explicitly models some available domain knowledge. We propose two instances of CRTM that incorporate prior knowledge in the form of document constraints. The models smooth the probability distribution of topics such that two constrained documents can either share the same topics or denote distinct themes. Experimental results on benchmark relational datasets show significant performances of CRTM on a semi-supervised document classification task.

Terragni, S., Fersini, E., Messina, E. (2020). Constrained Relational Topic Models. INFORMATION SCIENCES, 512, 581-594 [10.1016/j.ins.2019.09.039].

Constrained Relational Topic Models

Terragni, Silvia
Co-primo
;
Fersini, Elisabetta
Co-primo
;
Messina, Enza
Ultimo
2020

Abstract

Relational topic models (RTM) have been widely used to discover hidden topics in a collection of networked documents. In this paper, we introduce the class of Constrained Relational Topic Models (CRTM), a semi-supervised extension of RTM that, apart from modeling the structure of the document network, explicitly models some available domain knowledge. We propose two instances of CRTM that incorporate prior knowledge in the form of document constraints. The models smooth the probability distribution of topics such that two constrained documents can either share the same topics or denote distinct themes. Experimental results on benchmark relational datasets show significant performances of CRTM on a semi-supervised document classification task.
Articolo in rivista - Articolo scientifico
Constrained Relational Topic Models; Semi-supervised model; Latent Dirichlet Allocation; Domain knowledge
English
23-set-2019
2020
512
581
594
partially_open
Terragni, S., Fersini, E., Messina, E. (2020). Constrained Relational Topic Models. INFORMATION SCIENCES, 512, 581-594 [10.1016/j.ins.2019.09.039].
File in questo prodotto:
File Dimensione Formato  
Constrained Relational Topic Models.pdf

Solo gestori archivio

Tipologia di allegato: Publisher’s Version (Version of Record, VoR)
Dimensione 3.53 MB
Formato Adobe PDF
3.53 MB Adobe PDF   Visualizza/Apri   Richiedi una copia
Constrained_Relational_Topic_Models(1).pdf

accesso aperto

Tipologia di allegato: Submitted Version (Pre-print)
Dimensione 2.64 MB
Formato Adobe PDF
2.64 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/253986
Citazioni
  • Scopus 13
  • ???jsp.display-item.citation.isi??? 9
Social impact