Step Three: Define Model with Parameters and Filters
A new UI screen, AIML Training Template, has been introduced. This screen allows users to configure the parameters and details required to generate an AI/ML model for a specific metric.
The following fields are mandatory when creating a training template:
- Template Name
- AIML Metric
- Date Range
- AIML Algorithm
- The template name cannot contain a forward slash ( / ).
- Once saved, the template name cannot be edited.
In this screen, users can train and generate AI/ML models for future predictions. Similar to the Wave Template UI, the AIML Training Template screen allows users to configure specific parameters and settings required to generate a model for each supported metric.
The following KPIs are available for model creation:
- Order Cycle Time (regression problem)
- Order Waiting Time (regression problem)
- Order Picking Time (regression problem)
- Intelligent Cycle Count
Action Buttons Available in the Training Template UI
-
Train Template
This button is enabled only for metrics that support model generation (e.g., Order Cycle Time and Order Picking Time).
Once the template is saved, users can invoke the ‘Train Template’ action button. Upon clicking, the system will prompt the user to confirm whether they wish to proceed. If confirmed, the model training process begins, and the resulting model is captured and displayed in the AI/ML Models UI.
The ‘Predict’ action button is available only for specific metrics that do not require a trained model. For these KPIs, selecting Predict will directly generate output based on the available data—no model creation is necessary.
Users can define the date range for which they want to generate AI/ML KPI data. However, if the selected range does not contain sufficient data, the system will return the following error:
"Insufficient Training Data to Generate Model"
This screen supports full CRUD operations. When editing an existing template, all fields except the Template Name are editable. Once the required fields are properly configured with valid values, the template can be saved successfully.