Use this view to define the properties for the Time Series operator.
A Time Series operator is
a graphical icon representing a mining task that builds an associations
rule model that you place on the mining editor canvas.
In
the Properties view, set the properties for this operator by completing
the fields in the following tabs:
- General
- Model Name
- Column Properties
- Forecasting Settings
General tab
- Label
- You can rename the operator by specifying a new name. This new
name appears on the operator icon in the mining editor canvas.
- Description
- You can add a description for the operator. When the generated
model is stored in the IDMMX.TIMESERIESMODELS table,
the description is copied into the DESCRIPTION column of the table.
Model Name tab
- Prefix
- The prefix for the name of the model created by the operator.
Default
value: Name of the data warehousing project for
the mining flow
- Model name
- The name of the model that you want to build. The model is stored
in the IDMMX.TimeSeriesModels table. If a model with
the same name already exists, the previous model is replaced with
the new model. The maximum name length is 240.
Default value: System
determined
Column Properties tab
The tab is used to
specify which column of the input table is the time column that specifies
the order of the data set. And which columns contain the value fields
of the Time Series that are used for the forecast.
- Available columns
- This list contains all columns of the input table that are of
a numeric, date or time SQL type and can be used as Time or Value
columns. To specify them as time or value columns, select the column
you want to select and move them to the time or value columns sections
with the arrow buttons.
- Time Column
- The time column that specifies the order of the data set. You
have to select exactly one column. The column may not have duplicate
values.
- Value Columns
- The value columns that contain the values of the Time Series that
is to be forecast. You have to select at least one column. Note that
the Time Series visualizer can only compare two Time Series at a time.
Forecast Settings tab
This tab contains
settings for the forecasting algorithm. You can select the used algorithms,
the forecasting period and add optional parameters that are not available
as GUI setting.
- Forecasting Algorithms
- Three forecasting algorithms are supported:
- Seasonal Trend Decomposition
- Exponential Smoothing
- ARIMA
You can select the algorithms that are used for the forecasting.
All selected algorithms will be available in the Time Series model
and can be compared in the Time Series visualizer. Per default all
three algorithms are selected.
- Forecasting Period
- You can either use an automatic forecasting period or enter a
time point until which the forecasts are made. The default setting
for this parameter is Automatic Forecasting Period. With
this setting all algorithms will predict values until their internal
error prediction rises over a threshold. This results in forecasts
of different lengths for each algorithm since the error computation
of the algorithms is different. You can specify a fixed period until
which the algorithms shall predict the Time Series by selecting Manual
forecasting end point. The algorithms will predict the
Time Series until the specified point in time. If the end point is
too far into the future this will result in invalid forecasts and
huge confidence ranges.
- Optional Parameters
- This parameter is for advanced users only.
With optional parameter
strings, you can modify default parameters of the Easy Mining procedures.
For supported optional parameters refer to the Intelligent Miner® Easy Mining Procedures
documentation.
Example: DM_setPowerOptions('-buf 2000000')