Please enable JavaScript.
Coggle requires JavaScript to display documents.
CHAPTER 20: COMMUNICATE MODEL INSIGHTS (Information to be communicated…
CHAPTER 20: COMMUNICATE MODEL INSIGHTS
Information to be communicated during a presentation:
Business Problem
Model quality metrics (confusion matrix)
Areas where a model struggles
Most predictive features for model building
Feature types especially interesting to management
Recommended business actions
all algorithms work conceptually in the same way
determine the generalizable relationship between the features and the target
place those relationships into a model that can be used to understand those relationships and predict the outcome of cases not yet seen
Release/Unlock Holdout Sample
make sure you didn't make any mistakes in model creation process
if holdout sample scores are substantially lower than the cross-validation sample, we have reason to be worried about our modeling process
Look to see if business problem needs to be adjusted for new/additional information that has arisen during model process
Communicate confusion matrix
simplicity
animations showing movement between processes
explain what each term is (PPV, TPV, etc.) and what it means
Point out areas where the model struggles
Internal Data
External Data
Highlight most predictive features
FEATURES TO CONSIDER BEFORE GOING INTO PRESENTATION
Features that need to be changed and therefore requiring a re-run of the models
Features requiring further examination
subject matter expertise
Immutable features
good for modeling, no value to management if they want to implement change
Mutable features
features management could change
Recommend business actions