liveBook · Manning

Outlier detection

Also called novelty detection

Outlier detection is the process of identifying observations that differ substantially from expected or prevailing patterns in data.

How 3 Manning books explain it

Outlier Detection in Python

An outlier detection system is a system that identifies unusually divergent items in data and may provide scores or decisions about them.

How it is used
Finds unusual items for investigation across fraud, security, industrial monitoring, and data-quality work.
How deep it goes
Defines outliers through comparison, distinguishes learned detection from rules, and illustrates scoring without specifying a detector’s implementation.
Examples
Uses card purchases, network traffic, industrial sensor readings, and fund groupings to show different comparisons.
What you can do after
Evaluate/understand unusual-item flags as leads for contextual investigation rather than automatic findings of fraud, error, or failure.

Read how Outlier Detection in Python explains outlier detection →

Human-in-the-Loop Machine Learning: Active learning and annotation for human-centered AI

The identification of observations that differ substantially from the usual pattern in a dataset.

How it is used
Selects unfamiliar unlabeled items for human review to improve training-data diversity and expose gaps in a model’s knowledge.
How deep it goes
Develops an iterative word-match sampling procedure and discusses model-layer choices, including their biases and limitations.
Examples
Uses a rare disaster headline and a voice-assistant vocabulary problem to illustrate gaps that ordinary sampling can miss.
What you can do after
Apply outlier detection to assemble a human-review batch using word matches or model-layer representations.

Read how Human-in-the-Loop Machine Learning: Active learning and annotation for human-centered AI explains outlier detection →

How Large Language Models Work

Outlier detection is the identification of observations that are unusual relative to the prevailing pattern in data.

How it is used
Detects novel or unusual text requests and supports safer LLM solution design.
How deep it goes
Defines text dissimilarity, names two algorithms, and sketches embedding workflows, but does not derive detection mechanics.
Examples
Uses contrarian customers, novel problems, unusual requests, and customer-support comparisons with known solutions.
What you can do after
Diagnose unusual texts with embeddings and classic algorithms while retaining prevention or human escalation for uncertain cases.

Read how How Large Language Models Work explains outlier detection →

Also covered in

What to know first

Related concepts

Questions about outlier detection

What is outlier detection?

Outlier detection is the process of identifying observations that differ substantially from expected or prevailing patterns in data.

Which Manning books explain outlier detection?

Outlier Detection in Python, Human-in-the-Loop Machine Learning: Active learning and annotation for human-centered AI and How Large Language Models Work.

What should I know before learning outlier detection?

Manning books that teach outlier detection usually introduce clustering, correlation, histogram and language model first.

What kinds of outlier detection are there?

Kinds of outlier detection include cluster-based outlier detection, density-based outlier detection, frequent item set-based outlier detection and semisupervised outlier detection.

What is outlier detection also called?

Outlier detection is also called novelty detection.

These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.