Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • 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
Adaptiv relevansmodell med betingadsannolikhetshierarki
KTH, School of Computer Science and Communication (CSC).
2013 (Swedish)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesisAlternative title
Adaptive relevance model with conditionalprobability hierarchy (English)
Abstract [sv]

Syftet med detta examensarbete är att implementera

en adaptiv relevansmodell för att kunna analysera hur

bra denna presterar med den lilla mängd träningsdata

som finns tillgänglig. Resultatet bygger på den betingade

sannolikhetshierarki som är framtagen av Zhou

et al. Denna

relevansmodell implementeras mellan användargränssnittet

och Apache Solr, sökmotorn som användes vid studiens

genomförande. Genom att implementera på detta sätt

kommer både den systembaserade relevansen, från Apache

Solr, och den användarbaserade relevansen, från denna

relevansmodell, att bidra till den totala relevansen.

Den betingade sannolikhetshierarkin använder sig av

två andra modeller, oberoendemodellen och fullmodellen,

för att skatta relevansen hos ett dokument givet den

ställda sökfrågan. Oberoendemodellen gör antagandet att

söktermerna i sökfrågan är oberoende varandra men

förkastar informationen i ordföljden, medan fullmodellen

tar hänsyn till den information ordföljden ger.

Resultatet visar att relevansmodellen kan omordna

sökresultatet i

98% av fallen. Med rätt parametervärden

rankar den det sökta dokumentet högre i

84% av

fallen, jämfört med resultaten utan att använda denna

relevansmodell.

Abstract [en]

The purpose with this master thesis is to implement an

adaptive relevance model to be able to analyze how well it

performs with the small amount of training data available.

The results is based on the conditional probability hierarchy

developed by Zhou

et al. This relevance model is

implemented between the user interface and Apache Solr,

the search engine used at the studies implementation. By

implementing it in this way the system based relevance,

from Apache Solr, and the user based relevance, from

this relevance model, are both contributing to the overall

relevance.

The conditional probability hierarchy uses two other

models, the independent model and the full model, to

estimate the relevance for a document given the search

query. The independent model makes the assumption that

the search terms in the query are independent from each

other but discard the information given from the order of

the search terms, while the full model takes account for this

information.

The result shows that the relevance model can

rearrange the search results in

98% of the time, and with the

correct parameter settings it ranks the searched document

higher in

84% of the cases, compared to the search results

without using this relevance model.

Place, publisher, year, edition, pages
2013.
National Category
Computer Science
Identifiers
URN: urn:nbn:se:kth:diva-142354OAI: oai:DiVA.org:kth-142354DiVA: diva2:699763
Educational program
Master of Science in Engineering - Computer Science and Technology
Supervisors
Examiners
Available from: 2014-03-13 Created: 2014-02-28 Last updated: 2014-03-13Bibliographically approved

Open Access in DiVA

fulltext(751 kB)119 downloads
File information
File name FULLTEXT01.pdfFile size 751 kBChecksum SHA-512
f2ae28fe3aee5a7c6d06a32e2a027aaf52fa2b6d4b2e58feaa9079719901e5b1262632f35c93e3f59f4ebe330399879cb8428737cb9f771c5c9651e7bb894216
Type fulltextMimetype application/pdf

By organisation
School of Computer Science and Communication (CSC)
Computer Science

Search outside of DiVA

GoogleGoogle Scholar
Total: 119 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 71 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • 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