Features represent the core abstraction in Featureform. They serve as inputs to machine learning models, providing context or observations that the model leverages to make inferences. In practice, feature engineering often yields the highest return on investment for data scientists, significantly improving model performance and reliability. Features are employed in two primary contexts: building training sets and serving for inference.
dataset[[entity_col, value_col]]
. Optionally, you can set the variant. The type can take on one of the following values: ff.Int
, ff.Int32
, ff.Int64
, ff.Float32
, ff.Float64
, ff.Timestamp
, ff.String
. Since features are typically served for inference to your trained model, it’s essential to specify the inference store for materializing the feature.
Example:
dataset[[entity_col, value_col, timestamp_col]]
. Optionally, you can set the variant. The type can take on one of the following values: ff.Int
, ff.Int32
, ff.Int64
, ff.Float32
, ff.Float64
, ff.Timestamp
, ff.String
, ff.Bool
. Since features are typically served for inference to your trained model, it’s essential to specify the inference store for materializing the feature. To maintain point-in-time correctness, only the most recent entity-feature pair is retained in the inference store.
Example:
features
method to serve your features.
Example: