1
Motivation
2
Models and model families
3
The modelling process
4
Interface desirables
4.1
Desirables necessary for conceptual clarity
4.2
Desirables necessary for practicality
5
Proposed interface: model fitting
5.1
Fitting a
model
object
5.2
Objects in play
5.2.1
Model calibration
5.2.2
Hyperparameter space definition
5.2.3
Hyperparameter search
5.2.4
Model and model family objects
5.3
Model Instantiation
5.4
Model fitting
5.5
Prediction
5.6
Shortcut methods
6
Proposed interface: using models
6.1
Performance assessment
7
Using models: interactive mode
8
Using models: programmatic / automated mode
9
Extension to unsupervised learning
9.1
Invertible Mappings
9.2
Unsupervised Transformations That Are Not Maps
10
Best Practices
11
Existing interfaces
11.1
Scikit-Learn
11.2
caret
11.3
mlr
11.4
Idiomatic modelling in R
Some thoughts on modelling in R
Part 7
Using models: interactive mode
full fledged diagnostics
visual model comparison
in depth performance assessment (and visualizes)
small scale prediction