Sistema para el análisis de datos DBAnalyzer, una nueva versión
Cargando...
Archivos
Fecha
Autores
Armas Santos, Arazay
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas
Resumen
La detección de errores en los datos mediante el análisis de datos es fundamental en el proceso de limpieza. En esta fase se encontrarán varios tipos de errores y se podrá, según la taxonomía trazada, determinar de qué manera corregir los mismos.
La herramienta de detección de errores DBAnalyzer, realiza el perfil de los datos para, según medidas estadísticas, determinar, en el caso de valores de atributos individuales, errores potenciales, y además determina relaciones entre pares de atributos del mismo dominio cuyo incumplimiento debe ser analizado también como posible error. Teniendo en cuenta que pueden existir en los datos de diferentes dominios relaciones que no han sido descubiertas es que nuestro trabajo se ha encaminado a dotar al DBAnalyzer de la posibilidad de utilizar la técnica de minería de datos por reglas de asociación, en este caso reglas que involucren más de dos atributos con la utilización del algoritmo Apriori, para detectar otros tipos de errores que puedan aparecer en Sistemas de Bases de datos de nuestro entorno.
The detection of errors in the data by means of the analysis of data is essential in the cleaning process. During this phase, it will be found different types of errors and it will be able to, according to the traced taxonomy, determine the best way to correct them. DBAnalyzer program to find errors carries out the profile of the data for, according to statistical measures, to determine, in the case of individual attributes values, potential errors, and it also determines relationships between two pairs of attributes of the same domain whose nonfulfillment should be also analyzed as a possible error. Having in mind that, there might be in the different domains data relationships that have not been discovered yet, our work has been guided to provide the DBAnalyzer with the possibility of using the data mining technique based on association rules, and in this case, rules that involve more than two attributes with the use of the Apriori algorithm, to detect some other types of errors that might appear in Data bases Systems around us.
The detection of errors in the data by means of the analysis of data is essential in the cleaning process. During this phase, it will be found different types of errors and it will be able to, according to the traced taxonomy, determine the best way to correct them. DBAnalyzer program to find errors carries out the profile of the data for, according to statistical measures, to determine, in the case of individual attributes values, potential errors, and it also determines relationships between two pairs of attributes of the same domain whose nonfulfillment should be also analyzed as a possible error. Having in mind that, there might be in the different domains data relationships that have not been discovered yet, our work has been guided to provide the DBAnalyzer with the possibility of using the data mining technique based on association rules, and in this case, rules that involve more than two attributes with the use of the Apriori algorithm, to detect some other types of errors that might appear in Data bases Systems around us.