How deeply does Outlier Detection in Python cover stacked ensemble?

Explains score preparation, classifier training, probability scoring, cross-validation, and representative-label limitations without deriving classifier internals.

Stacked ensemble

As explained in Outlier Detection in Python

A predictive model that learns from the outputs of multiple component models or detectors, often using those outputs as input features.

A stacked ensemble combines scores from multiple outlier detectors with a classifier that learns a final outlier prediction or score. Unlike fixed weighting, it can account for detector correlations and cases where detectors succeed or fail differently. section 14.6.5

Build a stacked ensemble by:

Combining the score columns with ground-truth, hand-labeled, or pseudo-label data. section 14.6.5 section 15.4.1
Training a classifier, such as a Random Forest, and using its positive-class probability as the final outlier score. section 14.6.5
Cross-validating the model and checking that examples from both classes represent real data. section 14.6.5 section 15.4.1

Labeled detector scores

Create a dataframe containing detector scores and a Label column, then fit a Random Forest classifier to predict whether each record is of interest. Binary labels are supported; graded relevance scores are preferred. section 15.4.1

How Outlier Detection in Python covers it

liveBook manning.com

Read Outlier Detection in Python on liveBook Get the book at manning.com

Related concepts

classifieruses · A classifier is a predictive computational model that assigns inputs to one or more discrete classes or categories.
ensemblekind of · An ensemble is a technique that combines or considers outputs from multiple models or component methods collectively to produce or assess a joint result.
labeluses · A label is a value or annotation associated with an item that identifies its target category, outcome, or reference answer for learning or evaluation.
outlier detectionuses · Outlier detection is the process of identifying observations that substantially deviate from an expected or prevailing pattern in data.
anomaly scoreuses · An anomaly score is a numerical measure of how anomalous an observation or event is relative to a reference or expected pattern.

This explainer was generated by Manning's knowledge graph from the text of Outlier Detection in Python. The section links point to the parts of the book it draws on.