2005 ASemanticKernelToClassifyText

• (Basili et al., 2005) ⇒ Roberto Basili, Marco Cammisa, Alessandro Moschitti. (2005). “A Semantic Kernel to Classify Text with Very Few Training Examples.” In: Proceedings of the ICML 2005 Workshop on Learning in Web Search.

Subject Headings:

Cited By

2006

• Roberto Basili, Marco Cammisa, and Alessandro Moschitti. (2006). “A Semantic Kernel to Classify Texts with Very Few Training Examples.” In: Informatica, 30.

Quotes

Abstract

• Web-mediated access to distributed information is a complex problem. Before any learning can start, Web objects (e.g. texts) have to be detected and ¯ltered accurately. In this perspective, text categorization is a useful device to ¯lter out irrelevant evidence before other learning processes take place on huge sources of candidate information. The drawback is the need of a large number of training documents. One way to reduce such number relates to the use of more e®ective document similarities based on prior knowledge. Unfortunately, previous work has shown that such information (e.g. WordNet) causes the decrease of retrieval accuracy.
• In this paper we propose kernel functions to add prior knowledge to learning algorithms for document classi¯cation. Such kernels use a term similarity measure based on the WordNet hierarchy. The kernel trick is used to implement such space in a balanced and statistically coherent way. Cross-validation results show the bene¯t of the approach for the Support Vector Machines when few training examples are available.

References

---

,

	Author	volume	Date Value	title	type	journal	titleUrl	doi	note	year
2005 ASemanticKernelToClassifyText	Roberto Basili Alessandro Moschitti Marco Cammisa			A Semantic Kernel to Classify Text with Very Few Training Examples			http://dit.unitn.it/~moschitt/articles/ICML2005-ws.pdf