Batch Fit Time Series
The batch fitting process for time series data works nearly identically to the batch fitting process for other data sets. Additionally, it has the same limitations as fitting a time series to a single data set:
- Every data set must include at least six data points.
- Every data set must be contiguous in the worksheet - only values at the beginning or end of the series can be missing or blank.
- All sample values must fall in the range -1E+37 <= X <= +1E+37.
Just as with the batch fitting process for other data, the primary advantages of batch fitting time series data are efficiency and correlations. When running a batch fit for time series, the configuration is done one time and the fitting process is run for each data set selected and, if selected, correlation data can be generated during the process.
Similar to other batch fitting, a time series batch fit requires that every data set uses the same configurations such as time series models to test, the number of elements in the resulting data set, the fit statistic to be used in testing. This includes any manual transformations (Function, Detrend, and Deseasonalize) that may be set under the 'Data Transformations' section. If 'Auto Detect' is selected, every time series data set will be analyzed and have transformations applied as necessary; however, each time series data set will be evaluated for transformations individually, and each may have different transformations applied.
If different data sets require different configurations, those data sets must be fit separately.