Minería de subgrafos conexos frecuentes en colecciones de grafos etiquetados
Frequent connected subgraph (FCS) mining is an interesting problem in data mining with wide applications in real life. Examples of such applications appear in knowledge discovery from several sources as: chemical compound databases, XML documents, citation networks, biological networks, etc. For sol...
| Author: | |
|---|---|
| Format: | doctoral thesis |
| Status: | Versión aceptada para publicación |
| Publication Date: | 2010 |
| Country: | México |
| Institution: | Instituto Nacional de Astrofísica, Óptica y Electrónica |
| Repository: | Repositorio Institucional del INAOE |
| Language: | Spanish |
| OAI Identifier: | oai:inaoe.repositorioinstitucional.mx:1009/507 |
| Online Access: | http://inaoe.repositorioinstitucional.mx/jspui/handle/1009/507 |
| Access Level: | Open access |
| Keyword: | info:eu-repo/classification/Minería de datos/Data mining info:eu-repo/classification/Teoría de grafos/Graph theory info:eu-repo/classification/Teoría de algoritmos/Algorithm theory info:eu-repo/classification/cti/1 info:eu-repo/classification/cti/12 info:eu-repo/classification/cti/1203 |
| Summary: | Frequent connected subgraph (FCS) mining is an interesting problem in data mining with wide applications in real life. Examples of such applications appear in knowledge discovery from several sources as: chemical compound databases, XML documents, citation networks, biological networks, etc. For solving the FCS mining problem, several algorithms have been proposed to find all frequent connected subgraphs in collections of labeled graphs. The best performances in FCS mining are achieved by pattern growth-based algorithms. However, duplicate detection and support counting, during the mining process, are still a challenge in FCS mining. The four algorithms proposed in this thesis are focused on solving these subtasks in an efficient way regarding the reported methods. Specifically, novel theoretical properties of the candidate graph codifications are introduced. These properties allow to reduce the computational cost of duplicate detection in the practice. Moreover, a new data structure for indexing the graph collection is also introduced. This structure allows to accelerate support counting and candidate enumeration. On the other hand, there are some applications where analyzing all FCS is impractical. A way to face this problem is by finding only the closed FCS. Moreover, using the closed FCS set allows to reconstruct the whole set of FCS. |
|---|