Stretching the life of Twitter classifiers with time-stamped semantic graphs

Amparo Elizabeth Cano*, Yulan He, Harith Alani

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Social media has become an effective channel for communicating both trends and public opinion on current events. However the automatic topic classification of social media content pose various challenges. Topic classification is a common technique used for automatically capturing themes that emerge from social media streams. However, such techniques are sensitive to the evolution of topics when new event-dependent vocabularies start to emerge (e.g., Crimea becoming relevant to War Conflict during the Ukraine crisis in 2014). Therefore, traditional supervised classification methods which rely on labelled data could rapidly become outdated. In this paper we propose a novel transfer learning approach to address the classification task of new data when the only available labelled data belong to a previous epoch. This approach relies on the incorporation of knowledge from DBpedia graphs. Our findings show promising results in understanding how features age, and how semantic features can support the evolution of topic classifiers.

Original languageEnglish
Title of host publicationThe Semantic Web – ISWC 2014
Subtitle of host publication13th International Semantic Web Conference, Riva del Garda, Italy, October 19-23, 2014, proceedings, part II
EditorsPeter Mika, Tania Tudorache, Abraham Bernstein, Chris Welty, et al
Place of PublicationChem (CH)
PublisherSpringer
Pages341-357
Number of pages17
ISBN (Electronic)978-3-319-11915-1
ISBN (Print)978-3-319-11914-4
DOIs
Publication statusPublished - 31 Dec 2014
Event13th International Semantic Web Conference - Riva del Garda, Italy
Duration: 19 Oct 201423 Oct 2014

Publication series

NameLecture notes in computer science
PublisherSpringer
Volume8797
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference13th International Semantic Web Conference
Abbreviated titleISWC 2014
CountryItaly
CityRiva del Garda
Period19/10/1423/10/14

Fingerprint

Social Media
Stretching
Classifiers
Semantics
Classifier
Graph in graph theory
Transfer Learning
Supervised Classification
Dependent
Life

Keywords

  • concept drift
  • DBpedia
  • feature relevance decay
  • social media
  • topic detection

Cite this

Cano, A. E., He, Y., & Alani, H. (2014). Stretching the life of Twitter classifiers with time-stamped semantic graphs. In P. Mika, T. Tudorache, A. Bernstein, C. Welty, & et al (Eds.), The Semantic Web – ISWC 2014: 13th International Semantic Web Conference, Riva del Garda, Italy, October 19-23, 2014, proceedings, part II (pp. 341-357). (Lecture notes in computer science; Vol. 8797). Chem (CH): Springer. https://doi.org/10.1007/978-3-319-11915-1_22
Cano, Amparo Elizabeth ; He, Yulan ; Alani, Harith. / Stretching the life of Twitter classifiers with time-stamped semantic graphs. The Semantic Web – ISWC 2014: 13th International Semantic Web Conference, Riva del Garda, Italy, October 19-23, 2014, proceedings, part II. editor / Peter Mika ; Tania Tudorache ; Abraham Bernstein ; Chris Welty ; et al. Chem (CH) : Springer, 2014. pp. 341-357 (Lecture notes in computer science).
@inproceedings{30ced74fa5e241f09fad6974093c12c0,
title = "Stretching the life of Twitter classifiers with time-stamped semantic graphs",
abstract = "Social media has become an effective channel for communicating both trends and public opinion on current events. However the automatic topic classification of social media content pose various challenges. Topic classification is a common technique used for automatically capturing themes that emerge from social media streams. However, such techniques are sensitive to the evolution of topics when new event-dependent vocabularies start to emerge (e.g., Crimea becoming relevant to War Conflict during the Ukraine crisis in 2014). Therefore, traditional supervised classification methods which rely on labelled data could rapidly become outdated. In this paper we propose a novel transfer learning approach to address the classification task of new data when the only available labelled data belong to a previous epoch. This approach relies on the incorporation of knowledge from DBpedia graphs. Our findings show promising results in understanding how features age, and how semantic features can support the evolution of topic classifiers.",
keywords = "concept drift, DBpedia, feature relevance decay, social media, topic detection",
author = "Cano, {Amparo Elizabeth} and Yulan He and Harith Alani",
year = "2014",
month = "12",
day = "31",
doi = "10.1007/978-3-319-11915-1_22",
language = "English",
isbn = "978-3-319-11914-4",
series = "Lecture notes in computer science",
publisher = "Springer",
pages = "341--357",
editor = "Peter Mika and Tania Tudorache and Abraham Bernstein and Chris Welty and {et al}",
booktitle = "The Semantic Web – ISWC 2014",
address = "Germany",

}

Cano, AE, He, Y & Alani, H 2014, Stretching the life of Twitter classifiers with time-stamped semantic graphs. in P Mika, T Tudorache, A Bernstein, C Welty & et al (eds), The Semantic Web – ISWC 2014: 13th International Semantic Web Conference, Riva del Garda, Italy, October 19-23, 2014, proceedings, part II. Lecture notes in computer science, vol. 8797, Springer, Chem (CH), pp. 341-357, 13th International Semantic Web Conference , Riva del Garda, Italy, 19/10/14. https://doi.org/10.1007/978-3-319-11915-1_22

Stretching the life of Twitter classifiers with time-stamped semantic graphs. / Cano, Amparo Elizabeth; He, Yulan; Alani, Harith.

The Semantic Web – ISWC 2014: 13th International Semantic Web Conference, Riva del Garda, Italy, October 19-23, 2014, proceedings, part II. ed. / Peter Mika; Tania Tudorache; Abraham Bernstein; Chris Welty; et al. Chem (CH) : Springer, 2014. p. 341-357 (Lecture notes in computer science; Vol. 8797).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Stretching the life of Twitter classifiers with time-stamped semantic graphs

AU - Cano, Amparo Elizabeth

AU - He, Yulan

AU - Alani, Harith

PY - 2014/12/31

Y1 - 2014/12/31

N2 - Social media has become an effective channel for communicating both trends and public opinion on current events. However the automatic topic classification of social media content pose various challenges. Topic classification is a common technique used for automatically capturing themes that emerge from social media streams. However, such techniques are sensitive to the evolution of topics when new event-dependent vocabularies start to emerge (e.g., Crimea becoming relevant to War Conflict during the Ukraine crisis in 2014). Therefore, traditional supervised classification methods which rely on labelled data could rapidly become outdated. In this paper we propose a novel transfer learning approach to address the classification task of new data when the only available labelled data belong to a previous epoch. This approach relies on the incorporation of knowledge from DBpedia graphs. Our findings show promising results in understanding how features age, and how semantic features can support the evolution of topic classifiers.

AB - Social media has become an effective channel for communicating both trends and public opinion on current events. However the automatic topic classification of social media content pose various challenges. Topic classification is a common technique used for automatically capturing themes that emerge from social media streams. However, such techniques are sensitive to the evolution of topics when new event-dependent vocabularies start to emerge (e.g., Crimea becoming relevant to War Conflict during the Ukraine crisis in 2014). Therefore, traditional supervised classification methods which rely on labelled data could rapidly become outdated. In this paper we propose a novel transfer learning approach to address the classification task of new data when the only available labelled data belong to a previous epoch. This approach relies on the incorporation of knowledge from DBpedia graphs. Our findings show promising results in understanding how features age, and how semantic features can support the evolution of topic classifiers.

KW - concept drift

KW - DBpedia

KW - feature relevance decay

KW - social media

KW - topic detection

UR - http://www.scopus.com/inward/record.url?scp=84910093709&partnerID=8YFLogxK

UR - http://link.springer.com/chapter/10.1007/978-3-319-11915-1_22

U2 - 10.1007/978-3-319-11915-1_22

DO - 10.1007/978-3-319-11915-1_22

M3 - Conference contribution

AN - SCOPUS:84910093709

SN - 978-3-319-11914-4

T3 - Lecture notes in computer science

SP - 341

EP - 357

BT - The Semantic Web – ISWC 2014

A2 - Mika, Peter

A2 - Tudorache, Tania

A2 - Bernstein, Abraham

A2 - Welty, Chris

A2 - et al,

PB - Springer

CY - Chem (CH)

ER -

Cano AE, He Y, Alani H. Stretching the life of Twitter classifiers with time-stamped semantic graphs. In Mika P, Tudorache T, Bernstein A, Welty C, et al, editors, The Semantic Web – ISWC 2014: 13th International Semantic Web Conference, Riva del Garda, Italy, October 19-23, 2014, proceedings, part II. Chem (CH): Springer. 2014. p. 341-357. (Lecture notes in computer science). https://doi.org/10.1007/978-3-319-11915-1_22