umu.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Predicting User Competence from Text
Umeå University, Faculty of Science and Technology, Department of Computing Science. (Formal and Natural Language)
2017 (English)Conference paper, Published paper (Refereed)
Abstract [en]

We explore the possibility of learning user com-petence from a text by using natural language pro-cessing and machine learning (ML) methods. In ourcontext,competenceis defined as the ability to iden-tify the wildlife appearing in images and classifyinginto species correctly. We evaluate and compare theperformance (regarding accuracy and F-measure) ofthe three ML methods, Naive Bayes (NB), DecisionTrees (DT) and K-nearest neighbors (KNN), appliedto the text corpus obtained from the Snapshot Sen-rengeti discussion forum posts. The baseline resultsshow, that regarding accuracy, DT outperforms NBand KNN by 16.00%, and 15.00% respectively. Re-garding F-measure, K-NN outperforms NB and DTby 12.08% and 1.17%, respectively. We also proposeahybridmodelthatcombinesthethreemodels(DT, NB and KNN). We improve the baseline re-sults with the calibration technique and additionalfeatures. Adding a bi-gram feature has shown a dra-matic increase(from 48.38% to 64.40%) of accuracyfor NB model. We achieved to push the accuracylimit in the baseline models from 93.39% to 94.09%

Place, publisher, year, edition, pages
2017. 147-152 p.
Keyword [en]
text analysis, NLP, machine-learning, naive bayes, decision trees, and K-nearest neighbors
National Category
Language Technology (Computational Linguistics)
Research subject
Computer Science
Identifiers
URN: urn:nbn:se:umu:diva-138291OAI: oai:DiVA.org:umu-138291DiVA: diva2:1133833
Conference
21st World Multi-Conference on Systemics, Cybernetics and Informatics (WMSCI)
Available from: 2017-08-17 Created: 2017-08-17 Last updated: 2017-08-17

Open Access in DiVA

No full text

By organisation
Department of Computing Science
Language Technology (Computational Linguistics)

Search outside of DiVA

GoogleGoogle Scholar

Total: 9 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf