The ChannelMix Predictive Modeling Feature Metrics Dataset is used to power the Feature Metrics section of the Pipeline with Predictive Modeling Dashboard. The dataset contains metrics produced while training the model.
The dataset is refreshed at the same cadence as the Predictive Model is refreshed - monthly by default. Only the most recently produced results are displayed in the dashboard.
|Name of the feature in the model. I.e. Paid Social, Display, etc.
|The order a feature should appear in.
|The metric that the row is associated with. I.e. mROAS (marginal return on ad spend), weekly_seasonality, etc.
|Return Metric Value
|Importance of the feature in driving model predictions.
|The categorization of the row's feature. I.e. Paid, Seasonality
|Earliest date used to determine ROAS.
|Most recent date used to determine ROAS.
|Target Variable Pretty
A more descriptive name for the target variable to be displayed in the dashboard.
|The name of the field the model is trained to predict.
|The name of the OneView data is pulled from to model.
|The name of the feature used to predict the target variable.
|Name of the category in the model. I.e. Client 1, Client 2, Line of Business 1, Line of Business 2, etc.
|The name of the category used to create a hierarchy
|Model Run Date
|The date the model was trained.
|Training Start Date
|Earliest date provided to the model in training.
|Training End Date
|Most recent date provided to the model in training.
|Most Recent Model
|The date in which the record was loaded into the data warehouse
|The ChannelMix Profile associated to the data