EDOBE XDOM PMML Manual do Utilizador descarregar pdf (Página 17)

ARCHITECTURE FEATURES

The architecture for Spotfire Miner has a number of key features that

are important for understanding how the application works. This

section points out these features.

XML The product relies heavily on XML for storing and exchanging

information. XML is used as the storage format for worksheets,

model information, default settings, and other external files.

The initial implementation of Spotfire Miner also used Java XML

objects internally. These have been replaced with usage-specific data

structures for improved efficiency.

Pipeline Spotfire Miner’s computational engine is designed around a rich

pipeline architecture. The pipeline is a C++ infrastructure for passing

buffers of data between analytic components. The components

created in the graphical user interface map directly to related engine

computational components. This is discussed further in the section

The TIBCO Spotfire Pipeline on page 50.

Data Types The pipeline (and the product in general) knows about four data

types: continuous, categorical, string, and date.

Continuous

Continous columns are used to represent any sort of numerical data,

and are stored as double values.

Categorical

Categorical columns are used for values falling into a finite set of

categories, such as True/False or Small/Medium/Large. They are

stored as integer codes into a table of string labels. The string labels

are used for display.

String

String columns are used for informational columns such as names or

addresses that do not represent categories and are not used in

computations.

1 2 ... 12 13 14 15 16 17 18 19 20 21 22 ... 97 98

Sem comentários

EDOBE XDOM PMML Manual do Utilizador Página 17