TU Darmstadt / ULB / TUbiblio

Extrinsic Evaluation of Topic Models on Unknown Corpora

Schnober, Carsten ; Gurevych, Iryna (2015):
Extrinsic Evaluation of Topic Models on Unknown Corpora.
In: Expert Workshop on Topic Models and Corpus Analysis, pp. 1-2,
Dublin, DARIAH, Expert Workshop on Topic Models and Corpus Analysis, Dublin, Ireland, 14th - 15th December, 2015, [Conference or Workshop Item]

Abstract

We are presenting our practical experiences from using information retrieval (IR) as an extrinsic task for which, amongst many others, LDA can be applied (Wei et al., 2006). This allows us to use established evaluation metrics, precision and recall, in a realistic use case. We assume that a model that performs well on IR fits related tasks as well.

Item Type: Conference or Workshop Item
Erschienen: 2015
Creators: Schnober, Carsten ; Gurevych, Iryna
Title: Extrinsic Evaluation of Topic Models on Unknown Corpora
Language: English
Abstract:

We are presenting our practical experiences from using information retrieval (IR) as an extrinsic task for which, amongst many others, LDA can be applied (Wei et al., 2006). This allows us to use established evaluation metrics, precision and recall, in a realistic use case. We assume that a model that performs well on IR fits related tasks as well.

Book Title: Expert Workshop on Topic Models and Corpus Analysis
Place of Publication: Dublin
Publisher: DARIAH
Uncontrolled Keywords: reviewed;UKP_p_WeltDerKinder;UKP_reviewed;Topic Model, LDA, Corpus Exploration, Evaluation
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Ubiquitous Knowledge Processing
Event Title: Expert Workshop on Topic Models and Corpus Analysis
Event Location: Dublin, Ireland
Event Dates: 14th - 15th December, 2015
Date Deposited: 31 Dec 2016 14:29
URL / URN: https://dariah-tda.github.io/assets/cfp/evaluating_tms.pdf
Identification Number: TUD-CS-2015-1313
PPN:
Corresponding Links:
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details