05_introduction-to-instance-level-exploration.Rmd

# Introduction to Instance-level Exploration

**Learning objectives:**

- Define instance-level exploration methods
- Break-down plots
- Shapley additive explanations
- Local Interpretable Model-agnostic Explanations (LIME)
- Ceteris-paribus profiles and methods

## Instance-level exploration methods

Instance-level exploration methods help us understand how a model yields a prediction for a particular single observation.

- Evaluate effects of explanatory variables on the model’s predictions.
- How would the model’s predictions change if values of some of the explanatory variables changed? (what if analysis)
- Discover that the model is providing incorrect predictions, and we may want to find the reason.

## Break-down plots (variable attributions)

- Break-down plots for additive attributions (Chapter 6)
- Break-down plots for interactions (Chapter 7)
- Shapley additive explanations for average attributions (Chapter 8)

## Local Interpretable Model-agnostic Explanations (LIME)

- Uses the interpretation of the model as a function and investigates the local behavior of this function around the point (observation) of interest $x$
- "Local model" (Chapter 9)

## Ceteris-paribus profiles and methods

- *Ceteris paribus* is a Latin phrase that translates as "with other conditions remaining the same".
- Ceteris-paribus profiles show how a model's prediction would change if the value of a single explanatory variable changed. (Chapters 10 - 12)

## Meeting Videos {-}

### Cohort 1 {-}

`r knitr::include_url("https://www.youtube.com/embed/URL")`

<details>
<summary> Meeting chat log </summary>

```
LOG
```
</details>