Measures of similarity on graphs : Investigation and application to collaborative recommendation
(2007)
Files
No attached file found for this publication.
Details
- Authors
- Abstract
- (en) The amount of available information has been growing at a phenomenal rate, so that it is more and more difficult to process it. The challenge now consists in developing technologies that can help us sift through all this information. To uncover relationships in data, statistical techniques have been used for many years. Although traditional statistical techniques are still effective for problems involving small datasets and a manageable number of variables, their use make difficulties when applied to problems involving millions of records and thousands of variables. Data mining is thus emerging as a class of techniques enhancing statistics when examining large datasets. This work, suggesting various ways for computing similarities between nodes of a graph derived from the dataset, nicely becomes integrated into this data-mining framework. These algorithms are validated on their application to a personalization data-mining technology called collaborative recommendation.
- Affiliations
Louvain School of Management Operations and Information
Citations
Fouss, F. (2007). Measures of similarity on graphs : Investigation and application to collaborative recommendation. https://hdl.handle.net/2078.5/129925