mlr3 Manual
1
Quickstart
2
Introduction and Overview
3
mlr3 Basics
3.1
Quick R6 Intro for Beginners
3.2
Tasks
3.2.1
Task Types
3.2.2
Task Creation
3.2.3
Predefined tasks
3.2.4
Task API
3.2.4.1
Retrieving Data
3.2.4.2
Roles (Rows and Columns)
3.2.4.3
Task Mutators
3.3
Learners
3.3.1
Predefined Learners
3.4
Train & Predict
3.4.1
Basic concept
3.4.1.1
Creating Task and Learner Objects
3.4.1.2
Setting up the train/test splits of the data (#split-data)
3.4.1.3
Training the learner
3.4.1.4
Predicting
3.5
Resampling
3.5.1
Settings
3.5.2
Instantiation
3.5.3
Execution
3.5.4
Custom resampling
3.6
Benchmarking
3.6.1
Design Creation
3.6.2
Execution and Aggregation of Results
3.6.3
Converting specific benchmark objects to resample objects
3.7
Binary classification
3.7.1
ROC Curve and Thresholds
3.7.2
Threshold Tuning
4
Model Optimization
4.1
Hyperparameter Tuning
4.1.1
The
TuningInstance
Class
4.1.2
The
Tuner
Class
4.1.3
Triggering the Tuning
4.1.4
Automating the Tuning
4.2
Feature Selection / Filtering
4.2.1
Filters
4.2.2
Calculating filter values
4.2.3
Variable Importance Filters
4.2.4
Ensemble Methods
4.2.5
Wrapper Methods
4.3
Nested Resampling
4.3.1
Introduction
4.3.2
Execution
4.3.3
Evaluation
5
Pipelines
6
Technical
6.1
Parallelization
6.2
Error Handling
6.2.1
Encapsulation
6.2.2
Fallback learners
6.3
Database Backends
6.3.1
Example Data
6.3.2
Preprocessing with
dplyr
6.3.3
DataBackendDplyr
6.3.4
Model fitting
6.3.5
Cleanup
6.4
Parameters (using
paradox
)
6.4.1
Reference Based Objects
6.4.2
Defining a Parameter Space
6.4.2.1
Single Parameters
6.4.2.2
Parameter Sets
6.4.2.3
Vector Parameters
6.4.3
Parameter Sampling
6.4.3.1
Parameter Designs
6.4.3.2
Grid Design
6.4.3.3
Random Sampling
6.4.3.4
Generalized Sampling: The
Sampler
Class
6.4.4
Parameter Transformation
6.4.4.1
Transformation between Types
6.5
Logging and Verbosity
6.5.1
Available logging levels
6.5.2
Global Setting
6.5.3
Changing mlr3 logging levels
6.6
mlr -> mlr3 Transition Guide
7
Extending
7.1
Extending mlr3
7.1.1
Learners
7.1.1.1
Meta-information
7.1.1.2
Train function
7.1.1.3
Predict function
7.1.1.4
Final learner
7.2
Extending mlr3pipelines
8
Special Tasks
8.1
Survival Analysis
8.2
Spatial Analysis
8.3
Ordinal Analysis
8.4
Functional Analysis
8.4.1
How to model functional data?
8.5
Multilabel Classification
8.6
Cost-Sensitive Classification
8.6.1
A First Model
8.6.2
Cost-sensitive Measure
8.6.3
Thresholding
8.6.4
Threshold Tuning
9
Model Interpretation with mlr3
9.1
IML
9.2
Dalex
10
Use Cases
10.1
House Price Prediction in King County
10.1.1
Exploratory Data Analysis
10.1.1.1
Distribution of the price:
10.1.1.2
Association between variables
10.1.2
Splitting into train and test data
10.1.3
A first model: Decision Tree
10.1.4
A first baseline: Decision Tree
10.1.5
Many Trees: Random Forest
10.1.6
A better baseline:
AutoTuner
10.1.7
Engineering Features: Mutating ZIP-Codes
10.1.8
Obtaining a sparser model
11
Appendix
11.1
Integrated Learners
11.2
Integrated Filter Methods
11.2.1
Standalone filter methods
11.2.2
Algorithms With Embedded Filter Methods
References
Made with
for open-source
mlr3 manual
9.1
IML
