[ Home | Project Description | Partners | Papers | Prototypes | Restricted area ]
The Arianna Automatic CatalogThe Arianna Automatic Catalog is among the first web catalogs in the world realized in a completely automated manner. This page provides a short informative notice about the Automatic Catalog that has been realized as an engineered spin-off of the EuroSearch technology in the Arianna service. This page is organized as a FAQ, providing information about the innovative characteristics of the service.
What is an Automatic Catalog ?The incredible growth of the information available on the web will soon make impossible to maintain a catalog with the sole use of human resources, as it is currently done with traditional web catalogs. Based on this consideration, the automatic categorization technology has the objective to build a web catalog without human intervention, therefore allowing to:
By mean of abstracting techniques, even the site description is generated automatically from the most relevant text contained on the pages. What is the difference between an Automatic Catalog and a Search Engine ?An automatic catalog is different from a traditional search engine. However, it utilizes typical techniques of a search engine, such as spidering, indexing, searching and ranking, and adds specific techniques such as document clustering, classification and abstracting. While a search engine allows to search through an unstructured document space, an automatic catalog allows to navigate and search through a tree of thematic categories. What is the difference between an Automatic Catalog and a Traditional Catalog ?Compared with a traditional catalog, an automatic catalog is less subject to personal interpretation of the editors, is more frequently updated and analyses a bigger number of web sites. The use of the two types of catalogs is identical: both can be visited navigating through the categories tree or searching by keyword within one or more categories. How does the Automatic Catalog work ?The Arianna Automatic Catalog accesses directly the indexes of the Arianna search engine. At the heart of the system is the Automatic Classifier, which gets in input a set of category descriptions and applies specific algorithms for assigning documents to the best matching categories. Another important module is the Abstracter, which allows to extract the most relevant portions of text from the classified pages and build a short description of the web site. Who realized the Arianna Automatic Catalog ?The Arianna Automatic Catalog was partially developed within the EuroSearch project, and was born from a collaboration betwen Italia Online, the Department of Computer Science of the University of Pisa Medialab and I-Tech Studio di Ricerca. |