Explaining wine quality

Below is a SHAP visualization of the importance of data features used in a regression model to estimate the quality of red wine samples, using chemical properties (sulfates, alcohol, density, etc.) as predictor attributes or features.

  • The plot shows the relative importance of predictor attributes (chemical properties) in the inference process applied by the model for estimating the quality of a given observation (red wine sample).
  • The range of feature values is represented by colors ranging from blue (lowest) to pink (highest). The position of each feature's colored bar concerning the horizontal axis shows the effect feature values have in the resulting model output: some features' values lead to higher outputs when they are higher, while some other features' values lead to lower outputs when they are higher.

SHAP feature importance plot for wine quality prediction model

Take a deep look at the above SHAP plot, and select which of the following statements about model explainability and feature importance are true :

This exercise is part of the course

Understanding Artificial Intelligence

View Course

Hands-on interactive exercise

Turn theory into action with one of our interactive exercises

Start Exercise