Once a machine learning model has been deployed on the DataRobot server, this activity uses it to predict the target feature based on a single data set.
- DisplayName - The display name of the activity.
- Password - The password used to log into DataRobot.
- URL - (Optional) The URL of your DataRobot host. Defaults to https://app.datarobot.com if no value is provided.
- Username - The username used to log into DataRobot.
- Data - The local path of the data file used to predict the target feature. This file must contain tabular data (e.g. a .csv file) with each line missing a value for the target feature.
- Deployment - The name of the deployed model used to predict the target feature.
At design-time, enter your login credentials, hit "Load Deployments", and a list of deployed models will appear in the Predict activity!
Predictions - The predicted values of the target feature as a
Predictionobjects contain the following fields:
- Value (
String) - The predicted value for the model's target feature.
- Confidence (
Double) - The confidence, as a value between 0 and 1, that the predicted value is correct.
- PredictionExplanation (
Explanation[ ]) - An array of features used to train the model and the degree to which they influence the predicted value.
Explanationobjects contain the following fields:
- FeatureValue (
String) - The value this feature held in the input data.
- Strength (
Double) - A numerical indication of how strongly this feature correlated to the predicted value.
- Feature (
String) - The name of this feature.
- QualitativeStrength (
String) - A visual representation of how strongly this feature correlates with the predicted value. Possible values are ---, --, -, , +, ++, +++.
- Value (
Updated 3 months ago