A procedure to compute prototypes for data mining in non-structured domains
- First Online:
- Cite this paper as:
- Méndez J., Hernández M., Lorenzo J. (1998) A procedure to compute prototypes for data mining in non-structured domains. In: Żytkow J.M., Quafafou M. (eds) Principles of Data Mining and Knowledge Discovery. PKDD 1998. Lecture Notes in Computer Science, vol 1510. Springer, Berlin, Heidelberg
This paper describes a technique for associating a set of symbols with an event in the context of knowledge discovery in database or data mining. The set of symbols is related to the keywords in a database which is used as an implicit knowledge source. The aim of this approach is to discover the significant keyword groups which best represent the event. A significant contribution of this work is a procedure which obtains the representative prototype of a group of symbolic data. It can be used for both, unsupervised learning to describe classes, and supervised learning to compute prototypes. The procedure involves defining an objective function and the subsequent hypothesis-exploring system and obtaining an advantageous procedure regarding computational costs.
Key wordslearning data mining knowledge discovery symbolic clustering
Unable to display preview. Download preview PDF.