ML Modeller
Estimated reading: 7 minutes
230 views
Contents
Overview
Build powerful ML models in Infoveave for predictive analysis and decision-making. Configure filters, define input/output columns, set train-test ratios, and let Infoveave build the models. Save and explore the models for valuable insights and data-driven decision-making.
The machine learning models supported by Infoveave are:
- Binary Classifier: Make decisions by classifying observations as positive or negative.
- Regression: Predict changes in one variable based on historical data.
- Multi-Classifier: Classify problem instances into three or more classes for in-depth analysis.
- Clustering: Identify similar patterns or characteristics by grouping objects or cases into clusters.
Note
- Use the search option to look for any existing ML Model.
- Use the Filter by Datasource/Mode option to filter the ML Model based on the datasource.
- To switch the view between Card View and List View, click on the respective icons near to the search bar.
- To create a New Folder in the ML Model, click on the Folder icon.
- To download the list of all existing ML Model, click on the Download icon. It delivers the details on (• Entity Id Name • Description • Content Tags • Created By • Created By User • Created On • Folder • Datasource Id • Datasource Name • Is Folder Public)
Configure Machine Learning(ML) Model
Steps to configure ML model
To build an ML model in Infoveave, follow these steps:
Access the ML Modeller:
- Go to Analysis and select ML Modeller.
- Find your ML models under the tabs “My ML Models” and “Shared ML Models”.
Create a New ML Model:
- Click on “New Model” to start building a new ML model.
- Select the desired query from the dropdown list. For information on creating queries, refer to the relevant documentation.
- Choose the ML Model type from options like binary classifier, regression, multi-classifier, and clustering.
- Click “Save” to proceed and configure filters.

- Configure Filters:
- If required, set filters based on time progression, dimensions, and dimension items.
- For date progression filter, click on the “Order Date” field.
- To apply conditional filters, click on “Add Filter” and select a dimension from the dropdown list.
- Choose the appropriate conditional filters for the selected dimension (e.g., Exactly, Not Exactly, Contains, Doesn’t Contain).
- Delete applied conditional filters using the delete icon.
- If no filters can be applied to the selected query, Infoveave displays a message indicating so.
- Click on “Execute Query” to apply the filters and proceed to the ML Modeller window.

Note
- Note: Date dimensions offer the "Between" condition for selecting a date range
- We select Multi-Classifier Model here for the explanations.
- For the Clustering ML model, define the Cluster size.
- Model Configuration:
- In the ML Modeller window, define the Model Name String.
- Select the ML model type from the available options (binary classifier, regression, multi-classifier, and clustering) in the dropdown.
- If needed, modify the filters by clicking on “Filters“.
- Click on “View Table” to preview the data records after query execution.

- Column Configuration:
- The “Columns Info” section provides details about the columns involved in the ML model.
- Each column is defined by default as an input column with its respective data type.
- Click on a column to view or change its data type.
- Set each column as an input column, output column, or ignore column:
- Input column contributes to predicting the output column.
- Output column is the predicted column during model analysis.
- Ignore column is excluded from model building.
- Train-Test Ratio and Model Building:
- Select the Train-Test Ratio from the dropdown list (e.g., 70:30, 80:20, or custom).
- Click on “Build Model” after configuring columns, filters, and train-test ratio.
- Infoveave will start building the models and display them in the model info with accuracy macro.
- Save the Model:
- After building the models, select the desired model from the dropdown list.
- Click on “Save” and confirm the Predicted Column Name in the displayed dialog box.
- Click “Save” to create and view the model in the “My Model” tab.
Note
- In order to build the model, on you must define the output columns.
