Analytics
Score Model
Applies a model made by Linear Regression, Logistic Regression or Decision Tree to new rows, writing the prediction (or probability, or class) in a new column.
When to use it
Use it to score fresh data with a model fitted earlier — this month's customers with last month's churn model.
In a pipeline
File InputScore ModelExport CSV
Settings
| Setting | Type | Required | Default or hint |
|---|---|---|---|
| Model (JSON) | textarea | — | {"type":"linear","target":"Revenue","intercept":120,"coefficients":{"Spend":3.5}} |
| Or read the model from column | select | — | Optional: a Model column |
| Feature columns | select | — | Default: the columns the model names |
| Output | select | — | score |
| Score column name | text | — | Default: Score |
Example
Predict revenue for new accounts
- Model (JSON)={"type":"linear","target":"Revenue","features":["Spend"],"intercept":5,"coefficients":{"Spend":2}}
| Customer | Spend |
|---|---|
| New Co | 50 |
| Next Ltd | 70 |
| Customer | Spend | Score |
|---|---|---|
| New Co | 50 | 105 |
| Next Ltd | 70 | 145 |
Worth knowing
- Models are JSON, not stored artifacts: paste one from a model node's notes, or pick the Model column it added.
- The model's feature names are used as column names; pick Feature columns (one per feature, in order) when yours are named differently.
- Class output gives 1/0 for logistic models at the model's threshold, and the class for trees.