Feature interaction

As explained in Machine Learning System Design

A dependency or combined effect between features that influences their computation or a model's behavior.

Feature interactions combine signals or encode dependencies that change a model’s behavior and the way features are computed. In production, those dependencies determine calculation order, parallelization, serving-time data requirements, and whether a real-time SLA is achievable. section 11.1.1 section 11.1.2

Use feature interactions to expose relationships that individual features may hide, while checking their operational cost.

• Combine total rides and total distance into average distance per ride to capture driver or passenger behavior. section 11.1.2
• Compare product sales with other products in the same category or with earlier periods to reveal patterns and trends. section 11.1.2
• Prioritize interactions for linear models when combining features may make the data linearly separable. section 11.1.2
• Record dependencies and versions, then test and monitor interactions because fragile or unavailable inputs can destabilize predictions. section 11.1.1

Average distance per ride

A ride model may start with total rides and total distance. Dividing the latter by the former creates a combined signal that can describe usage more meaningfully than either absolute count alone. section 11.1.2

Figure 11.1 An overview of the DeepETA model pipeline: example of combining base feature engineering and a deep learning model (source: https://mng.bz/0MoN)

Use the pipeline view to see how base feature engineering can feed a deep learning model through combined signals. section 11.1.2

How Machine Learning System Design covers it

How deep it goes

Defines feature interactions, gives ratio and comparison examples, and notes linear separability; stops short of algorithmic construction details.

Examples

Uses average distance per ride, category and period sales comparisons, and a DeepETA feature-engineering pipeline.

What it assumes

Assumes existing features and training or inference pipelines; teaches how their signals can become derived features.

What you can do after

Build it from scratch by combining existing signals, then evaluate computational dependencies, serving availability, and maintenance risks.

How it connects

Connects feature engineering with linear models, deep learning, data pipelines, service-level agreements, and feature versioning.

liveBookmanning.com

Read Machine Learning System Design on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Machine Learning for Tabular Data: XGBoost, Deep Learning, and AIA feature interaction is a predictive effect produced jointly by two or more input features rather than by each feature independently.
• 

  liveBookmanning.com
  Ensemble Methods for Machine LearningA pairwise feature interaction is a joint effect in which a model's output depends on a combination of two features rather than either feature independently.
• 

  liveBookmanning.com
  Outlier Detection in PythonA feature interaction is a dependency in which the effect of one feature on a model output varies according to the value of another feature.

Compare how 4 books explain feature interaction →

Related concepts

• Explainable Boosted Machineused by · An Explainable Boosted Machine is an interpretable additive model trained with boosted shallow trees to learn individual feature effects and selected interactions.
• parallelismconfigures · Parallelism is the concurrent execution of multiple computational operations, often across multiple processing resources.
• tree-based methodaddressed by · A tree-based method is a predictive or decision-making technique that represents its rules or decision structure using one or more trees.
• linear regressionuses · Linear regression is a statistical modeling method that estimates a continuous response as a linear combination of predictor variables and fitted coefficients.
• linearly separableaddresses · Linearly separable describes labeled data whose classes can be divided by a single linear decision boundary in feature space.
• gradient boostinguses · Gradient boosting is an ensemble-learning technique that sequentially adds predictive models, typically weak learners, to reduce a loss by fitting its gradient.

This explainer was generated by Manning's knowledge graph from the text of Machine Learning System Design. The section links point to the parts of the book it draws on.