Filter Setup

You configure the following filters in order to filter out data you consider unacceptable for the calculation of the CDT. Note that the two attribute filters listed in Table 2-9 are stored and used during the Calculation stage. You also set additional data filtering parameters in the Calculation stage.

Figure 2-15 CDT Filter Setup

Description of "Figure 2-15 CDT Filter Setup"

Table 2-9 Data Filters

Data Filter Name	Data Filter Description
SKU Filter: Missing attribute values maximum	Each SKU is defined by its attribute values. If a certain absolute value for the attribute values is not defined, then the product definition is not accurate. A SKU with too many missing attribute values should be filtered out. The default value is 25 for the total attribute values (that is, a SKU with greater than 25 missing attribute values is not included in the calculation of the CDT).
Attribute Filter: Minimum attribute uses	An attribute that is used by only a few SKUs should be filtered out. The default value is an absolute value of 5 for the total SKUs in the category (that is, the data for an attribute that is used by fewer than five of the SKUs is not included in the calculation of the CDT). This filter does not remove t-log level data but instead removes the attribute and attribute values from the CDT creation process.
Attribute Value Filter: Minimum attribute value uses	An attribute value that is used by only a few SKUs should be filtered out. The default value is an absolute value of 5 for the SKUs in a category (that is, the data for an attribute value that is used by fewer than five of the SKUs is not included in the calculation of the CDT). This filter does not remove t-log level data but instead removes the attribute and attribute values from the CDT creation process.
Customer Filter: Transaction history minimum	Customers with short transaction histories are considered outliers. You assign a percentage value that is applied to the median number of transactions for all customers. Such customers are filtered out. The default value is 10% (that is, a customer who has fewer than 10% of the median number of transactions for all customers is not included in the calculation of the CDT).
SKU-Segment-Location Filter: Transaction minimum	SKUs that have few transactions for a given location-segment partition are considered outliers. You assign a percentage value that is applied to the median number of transactions for the SKUs in a specific partition. Such transactions are filtered out. The default value is 10% (that is, a SKU that is involved in fewer than 10% of the median number of transactions for a specific partition is not included in the calculation of the CDT).

Once you have configured the filters, click Run to start the filtering process.