|
Descripción
|
|
|---|---|
| Categorization of web search queries is of increasing interest not only due to increasing the effectiveness and efficiency of returned results but also for the potential revenue in coupled applications such us targeted advertising or reformulation of the query for better results. Nevertheless, categorization of the queries only based on the features of the query so far is challenging due to the small number of words in queries and the dynamism of sites and users requests. We define features of the queries based on visibility and decay of the terms they contain to be integrated with a taxonomy of concepts based on the site structure. The proposed method is basically composed of two steps: firstly some queries are categorized according to the results they return and a fast classifier is built based on these results. In a second stage, the method exploits properties of the visibility of the terms in the queries and in the front page along a period, to obtain a set of attributes that are used to further cluster queries and enrich classification. Altogether they integrate an online categorization of queries to help the site decision for targeted advertising. The classifier builder is based on Rough Set techniques integrated with traditional K-means and decision trees. Results of the classification on a site serving news and using GSA as search engine is shown | |
|
Internacional
|
Si |
|
Nombre congreso
|
ECML/PKDD |
|
Tipo de participación
|
960 |
|
Lugar del congreso
|
Varsovia, Polonia |
|
Revisores
|
Si |
|
ISBN o ISSN
|
|
|
DOI
|
|
|
Fecha inicio congreso
|
|
|
Fecha fin congreso
|
|
|
Desde la página
|
|
|
Hasta la página
|
|
|
Título de las actas
|
|