Use this stored procedure to print a model to get a summary
of its most important characteristics. Thus, you get a quick overview
of a model.
Authorities
The privileges held by the
authorization ID of the statement must include at least one of the
following authorities:
- SELECT privilege on the source model
- DATAACCESS authority
Syntax
IDAX.PRINT_MODEL(in parameter_string varchar(32672))
Parameter descriptions
- parameter_string
- Mandatory one-string parameter that contains pairs of
<parameter>=<value> entries that are separated by a comma.
- Data type: VARCHAR(32672)
- The following list shows the parameter values:
-
- model
- Mandatory.
- The name of the model that is to be printed.
- You can qualify the value of the model parameter with a schema name as
follows:
<schema name>.<model name>
Returned information
A printout as a result set.
The structure of the result set depends on the model type, that is, the algorithm, and the format
and contents of the model tables.
The following stored procedures have additional parameters and show specific returned
information. The algorithms are listed alphabetically.
Example
CALL IDAX.PRINT_MODEL('model=cc_nb_ewd');
Additional parameters for association rules models
- outtable
- Optional.
- The name of the model table to which the patterns and rules are written.
- Default: none
- Data type: VARCHAR(128)
- If no value is specified, the patterns or rules are only printed.
- type
- Optional.
- The type of information to display.
- Default:
rules
- Possible values are
rules
and patterns
.
- Data type: VARCHAR(8)
- limit
- Optional.
- The maximum number of rules or patterns that are displayed.
- Default: none
- Range: >=1
- Data type: INTEGER
- If no value is specified, all rules or patterns are displayed.
- itemsin
- Optional.
- A list of item names that must be contained in the rules or patterns that are displayed. The
items are separated by semicolons. At least one of the listed items must be contained.
- Default: none
- Data type: VARCHAR(32000)
- When the type=rules parameter is specified, the following conditions apply:
- To indicate that the item must be contained in the head of the rule, the item names can be
succeeded by :h or :head.
- To indicate that the item must be contained in the body of the rule, the item names can be
succeeded by :b or :body.
- If this parameter is not specified, no constraint is applied.
- itemsout
- Optional.
- A list of item names that must not be contained in the rules or patterns that are displayed. The
items are separated by semicolons.
- Default: none
- Data type: VARCHAR(32000)
- If this parameter is not specified, no constraint is applied.
- minlen
- Optional.
- The minimum number of items that are to be displayed in the rules or patterns.
- Default: 1
- Range: >= 1 and <= maxlen
- Data type: INTEGER
- maxlen
- Optional.
- The maximum number of items that are to be displayed in the rules or patterns.
- Default: the longest pattern of the model
- Range: >= minlen
- Data type: INTEGER
- minsupport
- Optional.
- The minimum support for items that are to be displayed in the rules or patterns.
- Default: 0.0
- Range: >=0.0 and <=maxsupport
- Data type: DOUBLE
- maxsupport
- Optional.
- The maximum support for items that are to be displayed in the rules or patterns.
- Default: 1.0
- Range: >=minsupport and <=1.0
- Data type: DOUBLE
- minlift
- Optional.
- The minimum lift of the rules or patterns that are to be displayed.
- Default: 0.0
- Range: >=0.0 and <=maxlift
- Data type: DOUBLE
- maxlift
- Optional.
- The maximum lift of the rules or patterns that are to be displayed.
- Default: the maximum lift of the patterns of the model
- Range: >=minlift
- Data type: DOUBLE
- minconf
- Optional.
- The minimum confidence of the rules that are to be displayed.
- Default: 0.0
- Data type: DOUBLE
- Range: >=0.0 and <= maxconf
- You can specify this parameter only if the type=rules parameter is
specified.
- maxconf
- Optional.
- The maximum confidence of the rules that are to be displayed.
- Default: 1.0
- Data type: DOUBLE
- Range: >=minconf and <= 1.0
- You can specify this parameter only if the type=rules parameter is
specified.
- sort
- Optional.
- A list of keywords that indicate the order in which the rules or patterns are displayed. The
order of the list is descending. The items are separated by semicolons.
- Possible values are
support
, confidence
, lift
, and length
.
- Default if the type=rules parameter is specified:
support;confidence;length
- Default if the type=patterns parameter is specified:
support;lift;length
- Data type: VARCHAR(32000)
Returned information for association rules
The result set contains all rules or patterns that meet the defined constraints in the order that
is specified in the sort parameter.
If the outtable parameter is specified, an output table that has the same
content as the result set is created.
Additional parameters for K-means models
- clusters
- Optional.
- The cluster IDs of the model that are to be printed.
- The cluster IDs are separated by a semicolon (;).
- If clusters is not specified, the cluster IDs for all clusters are
printed.
- columns
- Optional.
- The input column names of the model that are to be printed.
- The input column names are separated by a semicolon (;).
- If columns is not specified, the input column names for all input columns
are printed.
- mode
- Optional.
- The mode that determines which information of the model is to be printed.
- Allowed values are preview, clusters,
centers, statistics.
- If mode=preview is specified, the printout shows a table with one row for
each input column. The table has the columns mean (for numeric input columns) and mode (for nominal
input columns) for each cluster and for the overall data, and one row for the size of each
cluster.
- If mode=clusters is specified, the printout shows a table with one row per
cluster with the columns clusterid, name, size, withinss, and description.
- If mode=centers is specified, the printout shows a table with one row per
input column per cluster. The table has the columns clusterid, columnname, cardinality, mode,
minimum, maximum, mean, variance, count, and importance. If the model is built with
statistics=none, the table has the columns clusterid, columnname, mode (for
nominal input columns), and mean (for numeric input columns).
- If mode=statistics, the printout shows a table with one row per value of
the input columns and per cluster with the columns clusterid, columnname, value, count,
relfrequency, deviation, mean, and variance. If the model is built with
statistics=none or statistics=columns, the table is
empty.
Returned information for k-Nearest Neighbors (KNN)
The result set contains the column properties table that describes the ID, the target, and the
input columns of the model table. Statistics are collected for all input columns.
Table 1. Result set columns
| Column name |
Column type |
Description |
| COLNO |
INTEGER |
The rank of the column in the table |
| NAME |
VARCHAR(128) |
The name of the column |
| SQLTYPE |
VARCHAR(128) |
The SQL type of the column |
| LENGTH |
INTEGER |
The length of the SQL type of the column |
| SCALE |
SMALLINT |
The scale of the SQL type of the column |
| TYPE |
VARCHAR(64) |
The type of the column, that is, 'nom' or 'cont' |
| ROLE |
VARCHAR(64) |
The role of the column, that is, 'id', 'target', or 'input' |
| CARDINALITY |
BIGINT |
The number of distinct values of an input column |
| MINIMUM |
DOUBLE |
The minimum value of a continuous input column |
| MAXIMUM |
DOUBLE |
The maximum value of a continuous input column |
| MEAN |
DOUBLE |
The average value of a continuous input column |
| VARIANCE |
DOUBLE |
The variance of a continuous input column |
| NUMVALID |
BIGINT |
The number of valid values of an input column |
| NUMINVALID |
BIGINT |
The number of invalid values of an input column |
| NUMMISSING |
BIGINT |
The number of NULL values of an input column |
Returned information for linear regression
The result set depends on the specified options for the model creation.
- If the model is created with the calculatediagnostics parameter set to
false, one result set is returned.
- If the model is created with the calculatediagnostics parameter set to
true, two result sets are returned. The second result set contains additional diagnostic
information.
Table 2. Result set columns for result set 1
| Column name |
Column type |
Description |
| PREDICTOR |
VARCHAR Variable length that depends on the actual length of the predictor
names
|
The name of the predictor, that is, the column that is specified as the active input column
when the model is created Nominal columns can contain several entries for the same
predictor.
If the model is built with the intercept option set to true,
an additional predictor '(Intercept)' is included in the result set. This
predictor denotes the intercept that is included during the build process of the model.
|
| PREDICTOR_LEVEL |
VARCHAR Variable length that depends on the actual length of the predictor level
names
|
For nominal columns, the name of the predictor level, that is, the value of the column from
the input table for which the model is created If the predictor is
'(Intercept)' or of continuous data type, this column is intentionally left
empty.
|
| COEFFICIENT |
DOUBLE |
Coefficient of the predictor that is determined during model creation |
| STANDARD_DEVIATION |
DOUBLE |
Standard deviation of the predictor coefficient If the
calculatediagnostics parameter is set to false for the model creation, the
value is -1.
|
Table 3. Result set columns for result set 2 that contains additional diagnostic information
| Column name |
Column type |
Description |
| INDICATOR |
VARCHAR() |
Contains the following fixed values:
- [Y_VAR_EST]
- Estimators of a variance of the error term.
- [RSS]
- Residual sum of squares of the model.
- [R²]
- R-squared (coefficient of determination of the model).
|
| VALUE |
DOUBLE |
Indicator value |
Additional parameters for PCA models
- mode
- Optional.
- The mode that determines which information of the model is to be printed.
- Allowed values are standard and summary.
- Default: standard
Returned information for PCA
The result set depends on the specified options for the model creation.
For mode=standard, the output is as follows:
Table 4. Result set 1 – standard deviations vector for principal components
| Column name |
Column type |
Description |
| ID |
VARCHAR(3) |
Row index, always ‘[1] |
| PC<X> |
DOUBLE |
Standard deviation of principal component <X>, where X is a number of 1 <= n |
Table 5. Result set 2 – matrix of variable loadings
| Column name |
Column type |
Description |
| ID |
VARCHAR Variable length that depends on the name of the input variables
|
Name of the input variable |
| PC<X> |
DOUBLE |
Eigenvector value of the principal component <X>, where X is a number of 1 <= n |
For mode=summary, the output is as follows:
Table 6. Result set 1 - key statistics for principal components
| Column name |
Column type |
Description |
| MEASURE |
VARCHAR(27) |
Contains only the fixed values standard deviation, proportion of variance, and cumulative
proportion |
| PC<X> |
DOUBLE |
Indicator value for the principal component <X>, where X is a number of 1 <= n |
Additional parameters for sequential patterns
- outtable
- Optional.
- The name of the model table to which the patterns and rules are written.
- Default: none
- Data type: VARCHAR(128)
- If no value is specified, the patterns or rules are only printed.
- type
- Optional.
- The type of information to display.
- Default:
rules
- Possible values are
patterns
and rules
.
- Data type: VARCHAR(8)
- limit
- Optional.
- The maximum number of patterns or rules that are displayed.
- Default: none
- Range: >=1
- Data type: INTEGER
- If no value is specified, all rules or patterns are displayed.
- itemsin
- Optional.
- A list of item names that must be contained in the patterns or rules that are displayed. The
items are separated by semicolons. At least one of the listed items must be contained.
- Default: none
- Data type: VARCHAR(32000)
- When the type=rules parameter is specified, the following conditions apply:
- To indicate that the item must be contained in the head of the rule, the item names can be
succeeded by :h or :head.
- To indicate that the item must be contained in the body of the rule, the item names can be
succeeded by :b or :body
- If this parameter is not specified, no constraint is applied.
- itemsout
- Optional.
- A list of item names that must not be contained in the patterns or rules that are displayed. The
items are separated by semicolons.
- Default: none
- Data type: VARCHAR(32000)
- If this parameter is not specified, no constraint is applied.
- minlen
- Optional.
- The minimum number of items per pattern or rule that are to be displayed.
- Default: 1
- Range: >= 1 and <= maxlen
- Data type: INTEGER
- maxlen
- Optional.
- The maximum number of items per pattern or rule that are to be displayed.
- Default: the longest sequential pattern of the model
- Range: >= minlen
- Data type: INTEGER
- minsupport
- Optional.
- The minimum support of the patterns or rules that are to be displayed.
- Default: 0.0
- Range: >=0.0 and <=maxsupport
- Data type: DOUBLE
- maxsupport
- Optional.
- The maximum support of the patterns or rules that are to be displayed.
- Default: 1.0
- Range: >=minsupport and <=1.0
- Data type: DOUBLE
- minlift
- Optional.
- The minimum lift of the patterns or rules hat are to be displayed.
- Default: 0.0
- Range: >=0.0 and <=maxlift
- Data type: DOUBLE
- maxlift
- Optional.
- The minimum lift of the patterns or rules that are to be displayed.
- Default: the maximum lift of the sequential patterns of the model
- Range: >=minlift
- Data type: DOUBLE
- minconf
- Optional.
- The minimum confidence of the rules that are to be displayed.
- Default: 0.0
- Data type: DOUBLE
- Range: >=0.0 and <= maxconf
- You can specify this parameter only if the type=rules parameter is
specified.
- maxconf
- Optional.
- The maximum confidence of the rules that are to be displayed.
- Default: 1.0
- Data type: DOUBLE
- Range: >=minconf and <= 1.0
- You can specify this parameter only if the type=rules parameter is
specified.
- mintime
- Optional.
- The minimum time between the body and the head of the rules that are to be displayed.
- If the transaction ID of the data sequences on which the model is built is of type date, time or
time stamp, the value of this parameter is a number of seconds.
- Default: 0.0
- Data type: DOUBLE
- Range: >0.0 and <= maxtime
- You can specify this parameter only if the type=rules parameter is
specified.
- maxtime
- Optional.
- The maximum time between the body and the head of the rules that are displayed.
- If the transaction ID of the data sequences on which the model is built is of type date, time or
time stamp, the value of this parameter is a number of seconds.
- Default: The maximum relative time of the rules of the model
- Data type: DOUBLE
- Range: >=mintime
- You can specify this parameter only if the type=rules parameter is
specified.
- sort
- Optional.
- A list of keywords that indicate the order in which the rules or patterns are displayed. The
order of the list is descending. The items are separated by semicolons.
- Possible values are
support
, confidence
, lift
, length
, and
time
.
- Default if the type=rules parameter is specified:
support;confidence;length
- Default if the type=patterns parameter is specified:
support;lift;length
- Data type: VARCHAR(32000)
Returned information for sequential patterns
The result set contains all patterns or rules that meet the defined constraints in
the order that is specified in the sort parameter. If the
outtable parameter is specified, an output table that has the same content as
the textual representation is created.
Additional parameters for TwoStep clustering
- clusters
- Optional.
- The cluster IDs of the model that are to be printed.
- The cluster IDs are separated by a semicolon (;).
- If clusters is not specified, the cluster IDs for all clusters are
printed.
- Data type: VARCHAR(ANY)
- columns
- Optional.
- The input column names of the model that are to be printed.
- The input column names are separated by a semicolon (;).
- If columns is not specified, the input column names for all input columns
are printed.
- Data type: VARCHAR(ANY)
- mode
- Optional.
- The mode that determines which information of the model is to be printed.
- Allowed values are clusters, centers, and
statistics.
- Default: clusters
- Data type: VARCHAR(ANY)
Returned information for TwoStep clustering
Text.PrettyPrint of the TwoStep clustering model.