Hands-On Machine Learning with R Book Club
Welcome
Book club meetings
Pace
1
Introduction to Machine Learning
SLIDE 1
Meeting Videos
Cohort 1
2
Modeling Process
SLIDE 1
Meeting Videos
Cohort 1
3
Feature & Target Engineering
Introduction
3.2 Target engineering
3.3 Dealing with missingness
3.1
Missing values treatment
Imputation
3.4 Feature filtering
3.5 Numeric feature engineering
3.6 Categorical feature engineering
One-hot & dummy encoding
Alternatives
3.7 Dimension reduction
3.8 Proper implementation
Data leakage
Putting the pieces together
Meeting Videos
Cohort 1
4
Linear Regression
SLIDE 1
Meeting Videos
Cohort 1
5
Logistic Regression
SLIDE 1
Meeting Videos
Cohort 1
6
Regularized Regression
Introduction
6.1 - Why regularize?
6.2.1 - Ridge penalty
6.2.2 - Lasso penalty
6.2.3 - Elastic nets
6.3 - Implementation
6.4 - Tuning
6.5 - Feature interpretation
6.6 - Attrition data
6.7 - Final thoughts
Meeting Videos
Cohort 1
7
Multivariate Adaptive Regression Splines (MARS)
Extending linear models
Explaning the model
Loading prerequisites
7.0.1
Libraries to use
7.0.2
Data to use
Explaning model’s summary
Tuning Process
7.0.3
Caret
7.0.4
Tidymodels
Feature interpretation
Final thoughts
Meeting Videos
Cohort 1
8
K-Nearest Neighbors
SLIDE 1
Meeting Videos
Cohort 1
9
Decision Trees
9.1 - Introduction
9.2 - Structure
9.3 - Partitioning
9.4 - How deep?
9.4.1 - Early stopping
9.4.2 - Prunning
9.5 - AMES Housing example
9.6 - Feature Interpretation
9.7 - Final thoughts
9.1
BONUS: Attrition (decision tree classifier)
Meeting Videos
Cohort 1
10
Bagging (Bootstrap Aggregating)
It isn’t always useful
Out-of-bag (OOB) error
Prerequisites
10.0.1
Packages to load
10.0.2
Data to use
Training bagging trees
Number of trees and error estimate
Parallelize training
Error curve for custom parallel bagging
Feature interpretation
Final thoughts
Meeting Videos
Cohort 1
11
Random Forests
SLIDE 1
Meeting Videos
Cohort 1
12
Gradient Boosting
SLIDE 1
Meeting Videos
Cohort 1
13
Deep Learning
SLIDE 1
Meeting Videos
Cohort 1
14
Support Vector Machines
SLIDE 1
Meeting Videos
Cohort 1
15
Stacked Models
Base idea
Setting enviroment
15.1
Process Description
15.2
R packages available
15.3
Model Training Process
15.4
Alternative Model Training Process
15.5
Automated machine learning
Meeting Videos
Cohort 1
16
Interpretable Machine Learning
16.1
Introduction
16.2
Setting enviroment
16.2.1
Loading libraries
16.2.2
Getting the data
16.2.3
Training the model
16.2.4
Defining Local Observations to Explain
16.3
Interpretation trade-off
16.4
Model-specific implementation
16.5
Permutation-based feature importance
16.5.1
Concept
16.5.2
Implementation
16.6
Partial dependence
16.6.1
Concept
16.6.2
Implementation
16.6.3
Adding feature importance
16.7
Individual conditional expectation
16.7.1
Concept
16.7.2
Implementation
16.8
Feature interactions
16.8.1
Implementation
16.8.2
Alternatives
16.9
LIME
16.9.1
Implementation
16.10
Shapley values
16.10.1
Implementation
16.11
Localized step-wise procedure
16.11.1
Implementation
Meeting Videos
Cohort 1
17
Principal Components Analysis
SLIDE 1
Meeting Videos
Cohort 1
18
Generalized Low Rank Models
SLIDE 1
Meeting Videos
Cohort 1
19
Autoencoders
19.1
Main Concept
19.2
Applications
19.3
Prerequisites
19.4
Undercomplete autoencoders
19.4.1
Comparing PCA to an autoencoder
19.4.2
Coding example
19.4.3
Stacked autoencoders
19.4.4
Tuning hidden layers configuration
19.4.5
Anomaly detection
19.5
Denoising autoencoders
19.5.1
Corruption process
19.5.2
Coding example
19.6
Sparse autoencoders
19.6.1
Mathematical description
19.6.2
Tuning sparsity
\(\beta\)
parameter
19.7
Alternative autoencoders
19.8
Additinal references
Meeting Videos
Cohort 1
20
K-means Clustering
SLIDE 1
Meeting Videos
Cohort 1
21
Hierarchical Clustering
SLIDE 1
Meeting Videos
Cohort 1
22
Model-based Clustering
SLIDE 1
Meeting Videos
Cohort 1
Published with bookdown
Hands-On Machine Learning with R Book Club
Visualize the regression decision tree using
rpart.plot
.
rpart.plot
(ames_dt1)
