Machine Learning Model (MLM) tags enable you to deploy PMML (Predictive Model Markup Language) models converted from your previously created Python scripts. Your PMML models are used to create TrendMiner tags, that can be displayed and analyzed in TrendHub.
Note: This feature is only available if you have access management rights to the Machine Learning Model tags.
MLM submenu
PMML model: In this field you can select any of the models that are deployed on TrendMiner. The models can be deployed from the TrendMiner's Notebook tool, the platform where Python scripts can be converted into PMML models. Read here for more information on the Notebook tool.
Note: Only PMML models with 10 or less input variables will be shown in the dropdown field.
Mapping: Here you can map your tags/attributes to any required input for the model.
Note: Only analogue inputs are supported.
Output: Models can have multiple outputs. With this field you specify which output should be used to create your new tag.
Note: Only analogue outputs (integer, float, double) are supported.
Variables
A maximum of 10 variables can be included for mapping of each MLM tag. Only models with 10 or less input variables will be shown.
Input and Output
Only analogue inputs and outputs (integer, float, double) are supported.
Visualizing tags
Upon creating a new tag, the MLM tags are not immediately visible and may show up as blank data because the tag must first be indexed in TrendMiner. After indexing, the tag should be visible, you may have to remove and re-add the tag to your screen for this to happen.