Please enable JavaScript.
Coggle requires JavaScript to display documents.
Communicate Model Insights (5 types of info needed in a presentation…
Communicate Model Insights
Wahts imporatant os to distinguish between information that is useful for the model in advance
5 types of info needed in a presentation
Model Quality metrics (confusion metrics)
Ares where model struggles
Business Problem
Most predictive features for model building
Feature types especially interesting to MGMT
Recommended business actions
Unlocking Holdout
This checks to see if we made any mistakes]
Pick the best model to unlock holdout to create a data set with 100% of the training data used
we observe the corss-validation and holdout
Business Problem first
Pre-processing and Model Quality Metrics
the slides will provide a quick overview explaining the data and how we cleaned it
The model metrics and confusion matrix are combined to be better
Evaluating the most predictive featrues
Two main types of data
Internal data
Review the predictive ability
External data
mining public external data might be worthwile
4 kinds of features that need to be considered
features requiring further examination.
Immutable features.
Features that need to be changed and therefore requiring a re-run of the models
Mutable features.