When trying to use Rapid Deployment to deploy a Neural Network Time Series Model, you get the error message:

Neural Network Models that are supported for Rapid Deployment include SANN Regression, SANN Classificaiton and SANN Clustering. Other Time series type of SANN can not be deployed by current Rapid Deployment node.

Workaround:

In Interactive analysis mode:

1.  Save the PMML scripts after runing the SANN Time Series analysis

2.  When the testing/new data is active, select "Deploy models from previous analyses

3. Click "Load network files" to load previous saved PMML scripts for prediction on the testing/new data

Or before running the SANN analysis, configure the analysis to subset data into training sample and testing sample:

The training sample can be used to train the model, and the testing sample can be used to make new predictions.

1.Under Data Mining|Neural Networks, select the appropriate modeule

2. Select variables and strategy under the Quick Tab

3. Under Sampling (CNN and ANS) tab, select "Sampling variable:" as the Sampling method

4.Click "Training sample" to select the code for identifying training cases

5.Click "Testing sample" to select the code for identifying testing cases

In Workspace mode:

Configure Sampling(CNN and ANS) to subset data into Training sample and Testing sample