1 / 7

Simple, Proven Approaches To Text Retrieval

This presentation delves into simple and proven approaches to text retrieval, focusing on term matching and weighting strategies. Key concepts include collection frequency (CFW), term frequency (TF), document length (DL), and relevance weights. It discusses the importance of iterative searching and query expansion to enhance retrieval effectiveness. By combining evidence and using tuning constants, we can improve retrieval results significantly. Join us as we explore these methods and their implications for effective information retrieval.

rivka
Télécharger la présentation

Simple, Proven Approaches To Text Retrieval

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. Simple, Proven Approaches To Text Retrieval S. E. Robertson & K.Sparck Jones Presenters: Tuncer Turhan Yakup Korkmaz Ömer Köksal

  2. Term matching / weighting • Terms and matching • Sources for term weighting • Collectionfrequency n = the number of documents term t(i) occurs in N = the number of documents in the collection CFW (i) = log N - log n • Termfrequency TF (i,j) = the number of occurrences of term t(i) in document d(j)

  3. Term matching / weighting • Sources for term weighting (Continued) • Documentlength DL (j) = the total of term occurrences in document d(j) NDL (j) = (DL (j)) / (Average DL for all documents) • Combiningtheevidence CW (i,j) = [ CFW (i) * TF (i,j) * (K1+1) ] / [ K1 * ( (1-b) + (b * (NDL (j)) ) ) + TF (i,j) ] K1 and b are tuning constants.

  4. Iterative searching • Relevanceweights r = the number of known relevant documents term t(i) occurs in R = the number of known relevant document for a request RW (i) = log [ ( (r+0.5)(N-n-R+r+0.5) ) / ( (n-r+0.5)(R-r+0.5) ) ] • Queryexpansion OW (i) = r * RW (i)

  5. Iterative searching • Iterative combination CIW (i,j) = [ RW (i) * TF (i,j) * (K1+1) ] / [ K1 * ( (1-b) + (b * (NDL (j)) ) ) + TF (i,j) ]

  6. Details - Elaborations • Firstrequests • Longerqueries QACW (i) = QF(i) * CW(i,j) QACIW (i) = QF(i) * CIW(i,j) • Elaborations

  7. Thank you for listening…

More Related