1 / 26

Advanced best match

Advanced best match Mark Sanderson Porto, 2000 Aims To build on what you have done so far, reviewing more sophisticated ways of ranking documents in relation to a query. Objectives At the end of this lecture you will be able to describe a range of statistically based approaches to IR namely

albert
Télécharger la présentation

Advanced best match

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Advanced best match Mark Sanderson Porto, 2000

  2. Aims • To build on what you have done so far, reviewing more sophisticated ways of ranking documents in relation to a query.

  3. Objectives • At the end of this lecture you will be able to • describe a range of statistically based approaches to IR namely • Best match ranking • Passage retrieval • Web page popularity techniques • Web link counting

  4. Why? • Extend your knowledge in how IR systems work

  5. Up till now… • Inverse document frequency • More frequent - bad • Less frequent - good • More often a term is used (term frequency) • More likely document is about that term • Depends on document length?

  6. Basically covers it • More sophisticated formulas about • Basically tf*idf • Some variation, e.g. • tf normalisation • Divide by max tf? • Look at test collections • Singhal, A., Buckley, C., Mitra, M. (1996): Pivoted Document Length Normalization, in Proceedings of the 19th annual international ACM SIGIR conference on Research and development in information retrieval : 21-29

  7. Anything else? • How to make ranking better? • Term location? • Why? big documents • Web link analysis • Why? • Popularity • Why?

  8. Warning • Customise system to a collection… • …and it become less generally applicable.

  9. Term location • Prefer documents where terms are closer together? • Passage retrieval • Hearst, M.A., Plaunt, C. (1993): Subtopic structuring for full-length document access, in Proceedings of the 16th annual international ACM SIGIR conference on Research and Development in Information Retrieval: 59-68 • Callan, J. (1994): Passage­Level Evidence in Document Retrieval, in Proceedings of the 17th annual international ACM-SIGIR conference on Research and development in information retrieval: 302-310

  10. Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system searching over two years of the Financial Times newspaper. The system, called NRT (News Retrieval Tool), was described in a paper by Donna Harman, "User Friendly Systems Instead of User-Friendly Front Ends". Donna's paper appears in JASIS and the "Readings in IR" book by Sparck-Jones and Willett. 1st 2nd 3rd 4th Hearst and Plaunt • Split document into passages • Rank a document based on scores of each of it’s passages • What is a passage? • Paragraph? • TextTiling? • (non-overlapping) Fixed window?

  11. TextTiling? • Compute sentence block similarity

  12. Results? • TextTiling and paragraphs work equally well • Fixed windows work less well.

  13. Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system searching over two years of the Financial Times newspaper. The system, called NRT (News Retrieval Tool), was described in a paper by Donna Harman, "User Friendly Systems Instead of User-Friendly Front Ends". Donna's paper appears in JASIS and the "Readings in IR" book by Sparck-Jones and Willett. 1st 2nd 3rd 4th Callan • Split document into passages • Rank a document based on score of its highest ranking passage • What is a passage? • Paragraph? • Bounded paragraph • (overlapping) Fixed window?

  14. Results? • Overlapping passage did the best.

  15. Other use for best passage? • Anyone?

  16. Result list summaries

  17. How do they do that? • Best passage retrieval in the result list • Tombros, A. and Sanderson, M. Advantages of query-biased summaries in IR, in Proceedings of the 21st ACM SIGIR conference, Pages 2-10, 1998

  18. Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Web link analysis • Expand representation of document with text • Image retrieval by hypertext linksV. Harmandas, M. Sanderson and M.D. DunlopIn the Proceedings of the 20th ACM SIGIR conference, Pages 296-303, 1997

  19. Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Research interests/publications/lecturing/supervising My publications list probably does a reasonable job of ostensively defining my interests and past activities. For those with a preference for more explicit definitions, see below. I'm now working at the CIIR with an interest in automatically constructed categorisations and means of explaining these constructions to users. I also plan to work some more on a couple aspects of my thesis that look promising. Supervised by Keith van Rijsbergen in the Glasgow IR group, I finished my Ph.D. in 1997 looking at the issues surrounding the use of Word Sense Disambiguation applied to IR: a number of publications have resulted from this work. While doing my Ph.D., I was fortunate enough to apply for and get a small grant which enabled Ross Purves and I to investigate the use of IR ranked retrieval in the field of avalanche forecasting (snow avalanches that is), this resulted in a paper in JDoc. At Glasgow, I also worked on a number of TREC submissions and also co-wrote the guidelines for creating the very short queries introduced in TREC-6. Finally, I was involved in lecturing work on the AIS Advanced MSc course writing and presenting two short courses on Implementation and NLP applied to IR. I also supervised/co-supervised four MSc students. The work of three of these bright young things have been published in a number of good conferences. My first introduction to IR was the building (with Iain Campbell) of an interface to a probabilistic IR system Some sites better than others? • Hubs and authorities • Search for “Harvard” • Gibson, D., Kleinberg, J., Raghavan, P. (1998): Inferring Web Communities from Link Topology, in Proceedings of The 9th ACM Conference on Hypertext and Hypermedia: links, objects, time and space—structure in hypermedia systems: 225-234

  20. Like Google? • www.google.com • Base ranking on best match plus authority weight • Harder to spam authorities?

  21. Popularity? • Query IMDB (www.imdb.org) for “Titanic” Titanic (1915) Titanic (1997) Titanic (1943) Titanic (1953) Titanic 2000 (1999) Titanic: Anatomy of a Disaster (1997) Titanic: Answers from the Abyss (1999) Titanic Chronicles, The (1999) Titanic in a Tub: The Golden Age of Toy Boats (1981) Titanic Too: It Missed the Iceberg (2000) Titanic Town (1998) Titanic vals (1964)...aka Titanic Waltz (1964) (USA) Atlantic (1929)...aka Titanic: Disaster in the Atlantic (1999) (USA: video title) Night to Remember, A (1958)...aka Titanic latitudine 41 Nord (1958) (Italy) Gigantic (2000)...aka Untitled Titanic Spoof (1998) (USA: working title) …

  22. Use popularity • Query “titanic” on IMDB • Titanic (1997) • Titanic Too: It Missed the Iceberg (2000)

  23. Other examples • Direct Hit • www.directhit.com • Increasingly popular

  24. Figure out why yet?

  25. URL specific stuff • Prefer web pages with short URLs • http://dis.shef.ac.uk/mark/ • http://dis.shef.ac.uk/mark/trivia/links/stuff/ • Prefer with “.com” • not “.pt”?

  26. Conclusions • By now, you will be able to • describe a range of statistically based approaches to IR namely • Best match ranking • Passage retrieval • Web page popularity techniques • Web link counting

More Related