Non-Parametric Subject Prediction

Shenghui Wang; Rob Koopman; Gwenn Englebienne

doi:10.1007/978-3-030-30760-8_27

Non-Parametric Subject Prediction

Shenghui Wang^*, Rob Koopman, Gwenn Englebienne

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

12 Downloads (Pure)

Abstract

Automatic subject prediction is a desirable feature for modern digital library systems, as manual indexing can no longer cope with the rapid growth of digital collections. This is an “extreme multi-label classification” problem, where the objective is to assign a small subset of the most relevant subjects from an extremely large label set. Data sparsity and model scalability are the major challenges we need to address to solve it automatically. In this paper, we describe an efficient and effective embedding method that embeds terms, subjects and documents into the same semantic space, where similarity can be computed easily. We then propose a novel Non-Parametric Subject Prediction (NPSP) method and show how effectively it predicts even very specialised subjects, which are associated with few documents in the training set and are not predicted by state-of-the-art classifiers.

Original language	English
Title of host publication	Digital Libraries for Open Knowledge
Subtitle of host publication	23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019, Oslo, Norway, September 9-12, 2019, Proceedings
Editors	Antoine Doucet, Antoine Isaac, Koraljka Golub, Trond Aalberg, Adam Jatowt
Place of Publication	Cham
Publisher	Springer
Pages	312-326
Number of pages	15
ISBN (Electronic)	978-3-030-30760-8
ISBN (Print)	978-3-030-30759-2
DOIs	https://doi.org/10.1007/978-3-030-30760-8_27
Publication status	Published - 1 Jan 2019
Event	23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019 - OsloMet, Oslo, Norway Duration: 9 Sept 2019 → 12 Sept 2019 Conference number: 23 http://www.tpdl.eu/tpdl2019/contributions/

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Volume	11799
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019
Abbreviated title	TPDL 2019
Country/Territory	Norway
City	Oslo
Period	9/09/19 → 12/09/19
Internet address	http://www.tpdl.eu/tpdl2019/contributions/

Keywords

Non-parametric method
Random projection
Semantic embedding
Subject prediction

Access to Document

10.1007/978-3-030-30760-8_27

Cite this

Wang, S., Koopman, R., & Englebienne, G. (2019). Non-Parametric Subject Prediction. In A. Doucet, A. Isaac, K. Golub, T. Aalberg, & A. Jatowt (Eds.), Digital Libraries for Open Knowledge: 23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019, Oslo, Norway, September 9-12, 2019, Proceedings (pp. 312-326). (Lecture Notes in Computer Science; Vol. 11799). Springer. https://doi.org/10.1007/978-3-030-30760-8_27

Wang, Shenghui ; Koopman, Rob ; Englebienne, Gwenn. / Non-Parametric Subject Prediction. Digital Libraries for Open Knowledge: 23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019, Oslo, Norway, September 9-12, 2019, Proceedings. editor / Antoine Doucet ; Antoine Isaac ; Koraljka Golub ; Trond Aalberg ; Adam Jatowt. Cham : Springer, 2019. pp. 312-326 (Lecture Notes in Computer Science).

@inproceedings{ea8474aac3a04301a80efb39c8c48f7f,

title = "Non-Parametric Subject Prediction",

abstract = "Automatic subject prediction is a desirable feature for modern digital library systems, as manual indexing can no longer cope with the rapid growth of digital collections. This is an “extreme multi-label classification” problem, where the objective is to assign a small subset of the most relevant subjects from an extremely large label set. Data sparsity and model scalability are the major challenges we need to address to solve it automatically. In this paper, we describe an efficient and effective embedding method that embeds terms, subjects and documents into the same semantic space, where similarity can be computed easily. We then propose a novel Non-Parametric Subject Prediction (NPSP) method and show how effectively it predicts even very specialised subjects, which are associated with few documents in the training set and are not predicted by state-of-the-art classifiers.",

keywords = "Non-parametric method, Random projection, Semantic embedding, Subject prediction",

author = "Shenghui Wang and Rob Koopman and Gwenn Englebienne",

year = "2019",

month = jan,

day = "1",

doi = "10.1007/978-3-030-30760-8_27",

language = "English",

isbn = "978-3-030-30759-2",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "312--326",

editor = "Antoine Doucet and Antoine Isaac and Koraljka Golub and Trond Aalberg and Adam Jatowt",

booktitle = "Digital Libraries for Open Knowledge",

address = "Germany",

note = "23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019, TPDL 2019 ; Conference date: 09-09-2019 Through 12-09-2019",

url = "http://www.tpdl.eu/tpdl2019/contributions/",

}

Wang, S, Koopman, R & Englebienne, G 2019, Non-Parametric Subject Prediction. in A Doucet, A Isaac, K Golub, T Aalberg & A Jatowt (eds), Digital Libraries for Open Knowledge: 23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019, Oslo, Norway, September 9-12, 2019, Proceedings. Lecture Notes in Computer Science, vol. 11799, Springer, Cham, pp. 312-326, 23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019, Oslo, Norway, 9/09/19. https://doi.org/10.1007/978-3-030-30760-8_27

Non-Parametric Subject Prediction. / Wang, Shenghui; Koopman, Rob; Englebienne, Gwenn.
Digital Libraries for Open Knowledge: 23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019, Oslo, Norway, September 9-12, 2019, Proceedings. ed. / Antoine Doucet; Antoine Isaac; Koraljka Golub; Trond Aalberg; Adam Jatowt. Cham: Springer, 2019. p. 312-326 (Lecture Notes in Computer Science; Vol. 11799).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Non-Parametric Subject Prediction

AU - Wang, Shenghui

AU - Koopman, Rob

AU - Englebienne, Gwenn

N1 - Conference code: 23

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Automatic subject prediction is a desirable feature for modern digital library systems, as manual indexing can no longer cope with the rapid growth of digital collections. This is an “extreme multi-label classification” problem, where the objective is to assign a small subset of the most relevant subjects from an extremely large label set. Data sparsity and model scalability are the major challenges we need to address to solve it automatically. In this paper, we describe an efficient and effective embedding method that embeds terms, subjects and documents into the same semantic space, where similarity can be computed easily. We then propose a novel Non-Parametric Subject Prediction (NPSP) method and show how effectively it predicts even very specialised subjects, which are associated with few documents in the training set and are not predicted by state-of-the-art classifiers.

AB - Automatic subject prediction is a desirable feature for modern digital library systems, as manual indexing can no longer cope with the rapid growth of digital collections. This is an “extreme multi-label classification” problem, where the objective is to assign a small subset of the most relevant subjects from an extremely large label set. Data sparsity and model scalability are the major challenges we need to address to solve it automatically. In this paper, we describe an efficient and effective embedding method that embeds terms, subjects and documents into the same semantic space, where similarity can be computed easily. We then propose a novel Non-Parametric Subject Prediction (NPSP) method and show how effectively it predicts even very specialised subjects, which are associated with few documents in the training set and are not predicted by state-of-the-art classifiers.

KW - Non-parametric method

KW - Random projection

KW - Semantic embedding

KW - Subject prediction

UR - http://www.scopus.com/inward/record.url?scp=85072856969&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-30760-8_27

DO - 10.1007/978-3-030-30760-8_27

M3 - Conference contribution

AN - SCOPUS:85072856969

SN - 978-3-030-30759-2

T3 - Lecture Notes in Computer Science

SP - 312

EP - 326

BT - Digital Libraries for Open Knowledge

A2 - Doucet, Antoine

A2 - Isaac, Antoine

A2 - Golub, Koraljka

A2 - Aalberg, Trond

A2 - Jatowt, Adam

PB - Springer

CY - Cham

T2 - 23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019

Y2 - 9 September 2019 through 12 September 2019

ER -

Wang S, Koopman R, Englebienne G. Non-Parametric Subject Prediction. In Doucet A, Isaac A, Golub K, Aalberg T, Jatowt A, editors, Digital Libraries for Open Knowledge: 23rd International Conference on Theory and Practice of Digital Libraries, TPDL 2019, Oslo, Norway, September 9-12, 2019, Proceedings. Cham: Springer. 2019. p. 312-326. (Lecture Notes in Computer Science). doi: 10.1007/978-3-030-30760-8_27

Non-Parametric Subject Prediction

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this