# Machine Learning (ML) | Botable Glossary

https://www.botable.ai/glossary/machine-learning-ml

> Company-authored glossary / informational article — useful as an introductory reference on ML in QMS and regulatory expectations. Attribute claims to the site when reusing regulatory summaries and verify specific regulatory requirements against primary sources (e.g., FDA guidance) before treating them as definitive.

## Summary

The page defines Machine Learning (ML) in the context of Quality Management Systems (QMS) and medical devices, explains how ML can be used for predictive analytics and automation in QMS, and summarizes FDA regulatory expectations and validation requirements for ML-powered software and algorithm-driven medical devices.

## Audience

quality/regulatory and medical device professionals

## Prompts this page answers

- What regulatory requirements apply to machine-learning-powered medical device software?
- How can machine learning be used within a Quality Management System (QMS)?
- What dataset and validation requirements does the FDA expect for ML algorithms in medical devices?
- What examples of algorithm-driven devices require FDA special controls?

## Purpose

Educate readers about the role of machine learning in QMS and medical devices and outline regulatory and validation considerations for ML-powered software.

## Highlights

- ML is used in QMS and medical devices to analyze large datasets and predict potential quality issues.
- ML integration into medical devices triggers rigorous regulatory oversight, including requirements for software characterization, dataset descriptions, and independent validation.
- Verification and validation must include documentation of datasets used to train, tune, or test algorithms and use a validation dataset separate from training data.
- Labeling must describe algorithm limitations and failure modes and tests must demonstrate precision/reproducibility across inputs (operators, sites, devices).
- The FDA specifies special controls for various algorithm-driven devices (digital pathology, automated colony assessment, hemodynamic/predictive devices, radiological CAD) that require detailed algorithm documentation.

## How to cite

Botable Glossary — “Machine Learning (ML)” (link to the page)

## Publisher

**Botable** — Publisher of botable.ai.

## Topics

- machine learning in QMS
- ML medical devices regulation
- FDA software validation for algorithms
- predictive analytics IoT QMS
- algorithm validation dataset requirements

## Key entities

- **Botable** (organization): The website hosting the glossary entry. — https://www.botable.ai
- **FDA** (organization): Referenced regulator providing special controls and requirements for software and algorithm-driven devices.
- **Quality Management Systems (QMS)** (other): Context where ML is integrated, linked from the page. — https://www.botable.ai/teams/quality-regulatory
- **Internet of Things (IoT)** (other): Data source referenced as combined with ML for predictive analytics in QMS.

## Metadata

- Type: article