Posts

Cross validated predictions

Crossval predictions instead of the usual fit-predict approach.

Kernel visualization

Kernels can be seen as histogram generalizations.

Visualizing decision tree boundaries

Visualizing the decision intersections for a 2D classification via decision trees.

Isotonic regression

A lesser-known, step-like function approximation method.

Forests for feature importance

A classic example of using a (random) forest classifier to sort features.

Grid searching the optimal hyperparameter

Using sklearn GridSearch to optimize hyperparameters.

Topic modeling

Topic modeling with LDA.

kNN regression

Nearest neighbor regression with sklearn.

Feature selection with pipelines

Feature selection as part of a pipeline.

Naive Bayes

Gaussian naive Bayes classification basics.