GlOSS: text-source discovery over the Internet
read more
Citations
Routing indices for peer-to-peer systems
System for automatically generating queries
Database techniques for the World-Wide Web: a survey
Peer-to-peer information retrieval using self-organizing semantic overlay networks
System with user directed enrichment and import/export control
References
Introduction to Modern Information Retrieval
Automatic text processing: the transformation, analysis, and retrieval of information by computer
Searching distributed collections with inference networks
SIFT: a tool for wide-area information dissemination
Generalizing GlOSS to Vector-Space Databases and Broker Hierarchies
Related Papers (5)
Frequently Asked Questions (7)
Q2. How many real-user queries were used to test bGlOSS?
For test queries, the authors used a trace of 8,392 real-user queries issued at Stanford University to the INSPEC database from 4/12 to 4/25 in 1993.
Q3. What is the disadvantage of the hGlOSS server?
The hGlOSS server is very small in size and easily replicated, thus eliminating the potential bottleneck of the centralized GlOSS architecture.
Q4. How many entries are eliminated as threshold increases?
Adding all the indexes, the number of entries in the INSPEC frequency information kept by bGlOSS decreases very fast as threshold increases: for threshold51, for instance, 508,978 entries, or 46.71% of the original number of entries, are eliminated.
Q5. What is the weight of the words in the two documents that contain computer?
According to Assumption 2, each of the two documents that contain word computer will do so with weight 0.45 / 2 5 0.225, each of the 9 documents that contain word science will do so with weight 0.2 / 9 5 0.022, and so on.
Q6. What is the way to organize information in the Internet?
The Prospero File System is another example: Neuman [1992] lets users organize information in the Internet through the definition (and sharing) of customized views of the different objects and services.
Q7. What can be done to reduce the goodness of a query?
GlOSS can then use the negated terms to adjust the initial estimates, so that a database containing a negated term many times might see its goodness estimate for the query decreased.