MINEFE - Directions et services - Service de coordination à l’intelligence économique - La recherche et la veille sur le web. _predicting_the_present.pdf (application/pdf Objeto) An Empirical Analysis of Internet Search Engine Choice by Rahul Telang, Tridas Mukhopadhyay, Ronald Wilcox. Rahul Telang Carnegie Mellon University - H. John Heinz III School of Public Policy and Management Tridas Mukhopadhyay Carnegie Mellon University - David A. Ronald T. Carnegie Mellon UniversitySeptember 2001 Darden School of Business Working Paper No. 03-05 Abstract: We investigate consumers' choice behavior for Internet search engines.
Number of Pages in PDF File: 34 working papers series Suggested Citation Telang, Rahul and Mukhopadhyay, Tridas and Wilcox, Ronald T., An Empirical Analysis of Internet Search Engine Choice (September 2001). Using Search Engine for Classification: Does It Still Work? (Nik Corthaut) Geleijnse et al. [10] presented a very similar approach to [4] with similar results. We analyzedSchedls results in some detail, because we initiallyobtained far worse results, even with the same datasetand genre taxonomy. We only reached 50% accuracyversus the 62% reported by Schedl.
When we repeatedthe analysis, we obtained different results. Weanalyzed Schedls results and not Geleijnses, becauseof the larger data set. The topic of this paper is tovalidate whether this approach still works and analyzehow it performs on different search engines.First, we will briefly explain the approach, and thenwe will elaborate on the setup of our experiments and present the observations we made. 2.
For genre classification on artist level, Schedl et alrely on co-occurrence analysis. Pc . A ) to befound on a web page that mentions the genre name ( g )can be written formally as follows p(g|a) = pc a,g / pc with the page count of the artist and the pagecount of the combination of artist and genre. 3. 3.1.