Ejemplo 13: Analisis de Componentes Principales

Printer-friendly version

El cálculo de componentes principales se usa con frecuencia como un paso de procesamiento de la transformación de características. Puede reducir la dimensionalidad del conjunto de datos en cuestión, mientras se preservan las varianzas más importantes de los datos. Ejecutar el proceso y comprobar la salida en la vista gráfica del conjunto de datos Iris cargado y transformado por este proceso. 

1. Agregar el operador Repository Access → Retrieve a la zona de trabajo y localizar el archivo //Samples/data/Iris con el navegador del parámetro repository entry.

2. Agregar el operador Data Transformation → Value Modification → Numerical Value Modification → Normalize. Cambiar el nombre del mismo a “Normalización” y conectar la salida del operador Retrieve a la entrada exa (example set input) de este operador.

3. Agregar el operador Data Transformation → Attribute Set Reduction and Transformation → Principal Component Analysis. Cambiar el nombre del mismo a “Componentes Principales” y conectar la salida exa del operador Normalización (Normalize) a la entrada exa de este operador, y las salidas exa pre a sendos conectores res del panel.

 

 

Resultados:

 

 

Google
 
     

The best way to share the knowledge is to publish it in internet. Pues This is what go to achieve five libraries of Barcelona that take part in the project based in an agreement between the Generalitat of Catalonia and Google that consists in the digitalization of the impressive figure of more of 300.000 books. By if somebody is thinking already in the rights of author, the books that go to be...
Yesterday we had a new BI Beers in Barcelona , and beat attendance records. Otherwise I miscounted were 12, plus a girl who joined us for 15 minutes, until he realized we were talking about strange things and had the wrong group. As always, the conversation was very relaxed, and we share different view points on issues of BI. Frederic and Catherine were accompanied by more co- SAP , who...
  In a Data Warehouse there are a number of common dimensions such as Geographic and Time.  I leave here a script to create the Time Dimension table and a procedure to loading it between two dates:   / * Destination database * / use PAnalisys