SAP ABAP Data Element RSDMED_USE_WINDOWING (Use Sampling)

Hierarchy

☛ SAP_BW (Software Component) SAP Business Warehouse
   ⤷ BW-EI-DM (Application Component) Data Mining
     ⤷ RS_DME_DT (Package) Data mining engines: Decison trees

Basic Data

Data Element	RSDMED_USE_WINDOWING
Short Description	Use Sampling

Data Type

Category of Dictionary Type	D	Domain
Type of Object Referenced		No Information
Domain / Name of Reference Type	RSDMDM_CHAR_BOOL
Data Type	CHAR	Character String
Length	1
Decimal Places	0
Output Length	1
Value Table

Further Characteristics

Search Help: Name
Search Help: Parameters
Parameter ID
Default Component name
Change document	☐
No Input History	☐
Basic direction is set to LTR	☐
No BIDI Filtering	☐

Field Label

	Length	Field Label
Short	10	Sampling
Medium	15	Use Sampling
Long	23	Use Sampling Technique
Heading	36	Use Sampling Technique for Training

Documentation

Definition

Indicates whether to use sample data for training.

Use

To use the sampling technique in training, three parameters are required:

• Initial sample size
• Maximum size of data to be used
• Number of trials

In this technique, training is performed using sample data (equal to the specified initial sample size). The sample data contains an equal proportion of all class values. After training, the decision tree rules are applied to the remaining data.

Any misclassified data (up to a maximum of 20% of data size) is added to the sample data and training is performed again.

This continues till one of the following conditions is met:

• There are no misclassifications
• Tree accuracy is equal to or more than the specified accuracy (under stopping conditions)
• Training data size exceeds the specified maximum size

This process is repeated as many times as the number of trials specified.

Based on tree accuracy, the best tree is retained and all the other trees are discarded.

History

Last changed by/on	SAP	20110908
SAP Release Created in	30A