Data School |

Course outline: "Master Machine Learning with scikit-learn"

Course overview: "Master Machine Learning with scikit-learn"

Introduction to model ensembling

How to save a scikit-learn Pipeline with custom transformers

Should I shuffle samples with cross-validation?

Cost-sensitive learning in scikit-learn

scikit-learn vs Deep Learning

How to read the scikit-learn documentation

My top 50 scikit-learn tips

21 more pandas tricks

Adapt this pattern to solve many Machine Learning problems

Tune multiple models simultaneously with GridSearchCV

Access part of a Pipeline using slicing

Tune the parameters of a VotingClassifer or VotingRegressor

Ensemble multiple models using VotingClassifer or VotingRegressor

Create feature interactions using PolynomialFeatures

Speed up GridSearchCV using parallel processing

Use OrdinalEncoder instead of OneHotEncoder with tree-based models

Passthrough some columns and drop others in a ColumnTransformer

Drop the first category from binary features (only) with OneHotEncoder

Estimators only print parameters that have been changed

Load a toy dataset into a DataFrame

Get the feature names output by a ColumnTransformer

Create an interactive diagram of a Pipeline in Jupyter

Most parameters should be passed as keyword arguments

Don’t use .values when passing a pandas object to scikit-learn

Add feature selection to a Pipeline

Use FunctionTransformer to convert functions into transformers

Use AUC to evaluate multiclass problems

Shuffle your dataset when using cross_val_score

Four ways to examine the steps of a Pipeline

Vectorize two text columns in a ColumnTransformer

Save a model or Pipeline using joblib

Two ways to impute missing values for a categorical feature

Use stratified sampling with train_test_split

Prune a decision tree to avoid overfitting

Visualize a decision tree two different ways

Display the intercept and coefficients for a linear model

Use the correct methods for each type of Pipeline

Compare multiple ROC curves in a single plot

Plot a confusion matrix

Important tuning parameters for LogisticRegression

Display GridSearchCV or RandomizedSearchCV results in a DataFrame

Try RandomizedSearchCV if GridSearchCV is taking too long

Use cross_val_score and GridSearchCV on a Pipeline

Three reasons not to use drop=’first’ with OneHotEncoder

HistGradientBoostingClassifier natively supports missing values

Examine the intermediate steps in a Pipeline

What is the difference between Pipeline and make_pipeline?

Impute missing values using KNNImputer or IterativeImputer

Set a "random_state" to make your code reproducible

Add a missing indicator to encode "missingness" as a feature

Use Pipeline to chain together multiple steps

Handle unknown categories with OneHotEncoder by encoding them as zeros

Encode categorical features using OneHotEncoder or OrdinalEncoder

Four reasons to use scikit-learn (not pandas) for ML preprocessing

Use "fit_transform" on training data, but "transform" (only) on testing/new data

What is the difference between "fit" and "transform"?

Seven ways to select columns using ColumnTransformer

Use ColumnTransformer to apply different preprocessing to different columns