Descubrimiento de conjuntos frecuentes de ítems en datos estáticos y dinámicos
Currently the amount of data generated in any knowledge area is too big for being processed by a human. Among the more used data mining techniques are the mining or discovery of frequent item sets. In this thesis, two algorithms for frequent item sets (FI) mining on big sparse datasets are presented...
| Author: | |
|---|---|
| Format: | master thesis |
| Status: | Versión aceptada para publicación |
| Publication Date: | 2008 |
| Country: | México |
| Institution: | Instituto Nacional de Astrofísica, Óptica y Electrónica |
| Repository: | Repositorio Institucional del INAOE |
| Language: | Spanish |
| OAI Identifier: | oai:inaoe.repositorioinstitucional.mx:1009/829 |
| Online Access: | http://inaoe.repositorioinstitucional.mx/jspui/handle/1009/829 |
| Access Level: | Open access |
| Keyword: | info:eu-repo/classification/Minería de datos/Data mining info:eu-repo/classification/Reglas de asociación/Association rules info:eu-repo/classification/Conjuntos de datos dinámicos/Dynamic data sets info:eu-repo/classification/cti/1 info:eu-repo/classification/cti/12 info:eu-repo/classification/cti/1203 |
| Summary: | Currently the amount of data generated in any knowledge area is too big for being processed by a human. Among the more used data mining techniques are the mining or discovery of frequent item sets. In this thesis, two algorithms for frequent item sets (FI) mining on big sparse datasets are presented. The first algorithm named Compressed Arrays (CA) processes static data, i.e. dataset which do not change. Therefore, if the dataset is updated CA needs to process all the dataset to mine the new FI. CA performs a breadth first search through equivalence classes and introduces compressed arrays to accumulate the prefix class supports. CA is compared against the best algorithms reported in the literature. In our experiments, the best performance of CA algorithm was obtained for big sparse datasets. The second algorithm named Incremental Compressed Arrays (ICA) processes dynamic data, i.e. data in which a set of transactions can be added, deleted or modified. In order to mine the new FI after an updating, ICA does not need to process all the data but the current FI are used to obtain the new FI. Unlike previous algorithms, ICA does not suppose that the data fit in memory but it stores the mined FI in binary files. The experimentation shows than after adding, deleting or modifying a set of transactions, it is more efficient to use the FI previously mined than to process all the dataset from the beginning. |
|---|