A taxonomy of web search Andrei Broder IBM Research
The need behind a web search: - Informational - Navigational - Transactional Information queries constitute less than 50% of web searches The need behind a query might be : 2
Query Den klassiske model for IR Info Need Corpus (Samling) Search Engine Result Query Refinement 3
Den klassiske model for IR Task Info Need Verbal form QueryCorpus (Samling) Search Engine Result Query Refinement 4
1- Navigational queries Purpose: to reach a particular site that the user has in mind Eksempler: -American airlines home.(mulig mål: -Compaq. (mulig mål: -National car rental. ( mulig mål: Resultater som er en klik from målet kan være acceptabel, men mindre ønskelige. news.bbc.co.uk/ news.bbc.co.uk/hi/english/world/ Jubii: Google: Altavista: mange
2- Informational queries Purpose: to find informations, assumed to be avalable on the web in a static form IQ er closest to clasic IR 6
3- Transactional queries Purpose: to to reach a site where further interaction will happen -Shopping -Forskellige web – mediated servicer -Downloading forskellige type af filer( image, musik, osv.) -Adgang til bestemte database( nummeroplysning) -Finde servers 7
Statistic - Undersøgelse blandt AltaVistas brugere - Alalysis of the query log at AltaVista Which of the following describes best what you are trying to do? (navigational - non navigational) I want to get to a specific website that I already have in mind I want a good site on this topic, but I don’t have a specific site in mind Which of the following best describes why you conducted this search? I am shopping for something to buy on the Internet ( T ) I am shopping for something to buy elsewhere than on the Internet ( I ) I want to download a file (e.g., music, images, programs, etc.) None of these reasons Which of the following describes best what you are looking for? A site which is a collection of links to other sites regarding this topic The best site regarding this topic -24,53% - 68,41% - 8,16% - 5,46% - 22,55% - 57,19% - 14,83% -76,62% 8
Statistic - Undersøgelse blandt AltaVistas bruger - Alalysis of the query log at AltaVista Type of queriesUser SurveyQuery log Analysis Navigational24,5%20% InformationalEstimated: 39%48% TransactionalEstimated: 36%30% Queries with sexual content <1%Ca. 12% 9
The evolution of search engines: -First generation: on-page data, Informational queries -Second generation: use off-page, link analysis, informational –navigational -Third generation: analysis of need behind the query 10
Conclusion: -The need behind a query might be: -Informational -Navigational -Transactional -Search engines need to deal with all 3 -Current search engines deal well with IQ and NQ - 3.generations main aim is to effeciently with TQ, mostly via semantic analyses. understanding what the query is about). 11