MINING_SERVICES

Analysis Services Programming

previous page next page
Analysis Services Programming

MINING_SERVICES

This schema rowset provides a description of each data mining algorithm the provider supports.

Column name Type indicator
Description
SERVICE_NAME DBTYPE_WSTR The name of the algorithm. This column is provider-specific.
SERVICE_TYPE_ID DBTYPE_UI4 This column contains a bitmap that describes the mining service. Microsoft® SQL Server™ 2000 Analysis Services populates this column with one of the following values:
  • DM_SERVICETYPE_CLASSIFICATION (0x0000001)

  • DM_SERVICETYPE_CLUSTERING(0x0000002)
SERVICE_DISPLAY_NAME DBTYPE_WSTR A localizable display name for the algorithm.
SERVICE_GUID DBTYPE_GUID The GUID for the algorithm.
DESCRIPTION DBTYPE_WSTR A user-friendly description of the algorithm.
PREDICTION_LIMIT DBTYPE_UI4 The maximum number of predictions the model and algorithm can provide.
SUPPORTED_DISTRIBUTION_FLAGS DBTYPE_WSTR A comma-delimited list of flags that describe the statistical distributions supported by the algorithm. This column contains one or more of the following values:
  • "NORMAL"

  • "LOG NORMAL"

  • "UNIFORM"

Provider-specific flags can also be defined.
SUPPORTED_INPUT_CONTENT_TYPES DBTYPE_WSTR A comma-delimited list of flags that describe the input content types that are supported by the algorithm. This column contains one or more of the following values:
  • "KEY"

  • "DISCRETE"

  • "CONTINUOUS"

  • "DISCRETIZED"

  • "ORDERED"

  • "SEQUENCE TIME"

  • "CYCLICAL"

  • "PROBABILITY"

  • "VARIANCE"

  • "STDEV"

  • "SUPPORT"

  • "PROBABILITY VARIANCE"

  • "PROBABILITY STDEV"

  • "ORDER"

Provider-specific flags can also be defined.
SUPPORTED_PREDICTION_CONTENT_TYPES DBTYPE_WSTR A comma-delimited list of flags that describe the prediction content types that are supported by the algorithm. This column contains one or more of the following values:
  • "KEY"

  • "DISCRETE"

  • "CONTINUOUS"

  • "DISCRETIZED"

  • "ORDERED"

  • "SEQUENCE TIME"

  • "CYCLICAL"

  • "PROBABILITY"

  • "VARIANCE"

  • "STDEV"

  • "SUPPORT"

  • "PROBABILITY VARIANCE"

  • "PROBABILITY STDEV"

Provider-specific flags can also be defined.
SUPPORTED_MODELING_FLAGS DBTYPE_WSTR A comma-delimited list of the modeling flags that are supported by the algorithm. This column contains one or more of the following values:
  • "MODEL_EXISTENCE_ONLY"

  • "NOT NULL"

Provider-specific flags can also be defined.
SUPPORTED_SOURCE_QUERY DBTYPE_WSTR The <source_data_query> types that the provider supports. This is a comma-delimited list of one or more of the following syntax descriptions. These descriptions can be used as the source of data for INSERT INTO, or can be joined using a PREDICTION JOIN to a mining model for SELECT. The following values are available:
  • "SINGLETON CONSTANT"

  • "SINGLETON SELECT"

  • "OPENROWSET"

  • "SELECT"

  • "SHAPE"
TRAINING_COMPLEXITY DBTYPE_I4 Indicates the length of time training is expected to take:
  • DM_TRAINING_COMPLEXITY_LOW

    Running time is relatively short, and it is proportional to input.

  • DM_TRAINING_COMPLEXITY_MEDIUM

    Running time may be long, but it is generally proportional to input.

  • DM_TRAINING_COMPLEXITY_HIGH

    Running time is long and it may grow exponentially in relationship to the number of training cases.
PREDICTION_COMPLEXITY DBTYPE_I4 Indicates the length of time prediction is expected to take:
  • DM_PREDICTION_COMPLEXITY_LOW

    Running time is relatively short, and it is proportional to input.

  • DM_PREDICTION_COMPLEXITY_MEDIUM

    Running time may be long, but it is generally proportional to input.

  • DM_PREDICTION_COMPLEXITY_HIGH

    Running time is long and it may grow exponentially in relationship to the number of training cases.
EXPECTED_QUALITY DBTYPE_I4 Indicates the expected quality the of model produced with this algorithm:
  • DM_EXPECTED_QUALITY_LOW

  • DM_EXPECTED_QUALITY_MEDIUM

  • DM_EXPECTED_QUALITY_HIGH
SCALING DBTYPE_I4 Indicates the scalability of the algorithm:
  • DM_SCALING_LOW

  • DM_SCALING_MEDIUM

  • DM_SCALING_HIGH
ALLOW_INCREMENTAL_INSERT DBTYPE_BOOL Contains VARIANT_TRUE if additional INSERT INTO statements are allowed after the initial training.
ALLOW_PMML_INITIALIZATION DBTYPE_BOOL Contains VARIANT_TRUE if the mining models (including structure and content) can be created based on an XML string.
CONTROL DBTYPE_I4 Contains one of the following values, which determine whether the service supports training interruption:
  • DM_CONTROL_NONE

    The algorithm cannot be canceled after it starts to train the model.

  • DM_CONTROL_CANCEL

    The algorithm can be canceled after it starts to train the model, but must be restarted to resume training.

  • DM_CONTROL_SUSPENDRESUME

    The algorithm can be canceled and resumed at any time, but results are not available until training is complete.

  • DM_CONTROL_SUSPENDWITHRESULT

    The algorithm can be canceled and resumed at any time, and any incremental results can be obtained.
ALLOW_DUPLICATE_KEY DBTYPE_BOOL Contains VARIANT_TRUE if cases are allowed to contain duplicate keys.

Default Sort Order

SERVICE_NAME

Restriction Columns

SERVICE_NAME
SERVICE_TYPE_ID

previous page start next page