Font Size: a A A

The Research On Model Of Knowledge Discovery Based On Granularity And Concept Lattice

Posted on:2007-06-27Degree:MasterType:Thesis
Country:ChinaCandidate:L ZhangFull Text:PDF
GTID:2178360182986280Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
While processing the large amounts of data, the problem can be observed, analyzed from different conceptual level, different levels of granularity. Gaining different levels of information from different granularity of knowledge meets people's requirement of different levels of abstract knowledge. Concept Lattice is a powerful model for data analysis and rule extraction, which is one of the most important models on data description. While the existing models based on concept lattice aim at the static database and the attribute value is just from the single level, lacking of the transformation of granularity.In this thesis, the granularity is introduced to concept lattice model. Combined with the granularity transformation and the knowledge of concept lattice, the existing data mining methods based on concept lattice are extended. The contexts of the thesis are as follows:1. The basic notion and the background of the data mining and KDD, the mathematical foundation, the concept lattice model and the extend concept lattice model are introduced. The features of various algorithms which used for building the lattice are discussed. The three major models for granular computing are researched after discussing the basic components and issues.2. In order to accommodate to the transformation of granularity of the attribute value, the dynamic transform algorithm of concept lattice for multi-granularity attribute values is proposed when considering the change of nodes as a result of transformation the granularity have been analyzed. The experiments show that the dynamic transform algorithm is superior to the algorithm which rebuilds the lattice.3. Then the context of the multi-granularity attribute Values is proposed for the requirement of the transformation of granularity. Still since the nodes of concept lattice can not store and express the granularity of the attribute values, how to store and express it will be discussed. The KDD model based on the granularity and concept lattice is proposed at last.
Keywords/Search Tags:Data Mining, Concept Lattice, Granularity, Granular Computing
PDF Full Text Request
Related items