AI/ML glossary

AI and machine learning terms, explained for marketers

Plain-language definitions of 283 terms from machine learning, NLP and AI search, each with how it is used in SEO and marketing.

Entities & Knowledge Graphs
@id (Schema Node Identifier)
@id is the JSON-LD property that gives an entity in your structured data a unique, stable identifier, so every page can refer to the same entity.
Model Evaluation
Accuracy (ML metric)
Accuracy is the share of all predictions a model gets right. It is the simplest evaluation metric, but it misleads when one class is much more common.
Learning Paradigms
Active Learning
A training approach where the model picks the unlabelled examples it is least sure about for a human to label, so you reach good accuracy with fewer l…
AI Search & Agents
Agentic AI
AI systems that pursue a goal over several steps on their own: planning, choosing tools or sources, acting and checking results with little human inpu…
AI Search & Agents
Agentic Commerce Protocol (ACP)
An open standard from OpenAI and Stripe defining how AI agents complete purchases with businesses on a buyer's behalf, using scoped payment tokens.
AI Search & Agents
Agentic Search
Search where an AI system plans, runs and evaluates many searches itself, then delivers one synthesised answer instead of a list of links.
AI Search & Agents
AI Agent
A system that pursues a goal on its own: it plans, uses tools, keeps track of what it has learned and adapts until the task is done.
AI Search & Agents
AI Citation (LLM Citation)
A source link or named mention an AI search system attaches to its answer, showing which pages or brands it drew on and credited.
AI Search & Agents
AI Crawlers (GPTBot, ClaudeBot, PerplexityBot)
Bots run by AI companies that fetch web pages for model training, AI search indexes or live answers; you control most of them in robots.txt.
Generative AI & LLMs
AI Governance
The policies, roles and controls an organisation uses to make sure AI is used safely, legally, transparently and in line with its goals.
Generative AI & LLMs
AI Hallucination
When a generative AI model produces content that sounds plausible but is false, unsupported by its sources, or invented outright.
AI Search & Agents
AI Mode (Google)
Google AI Mode is a conversational search experience that answers with a Gemini-written response, built by fanning one question out into many searches…
AI Search & Agents
AI Overview
Google's AI-generated answer shown above some search results, summarising information from several pages and linking to supporting sources.
AI Search & Agents
AI Rank Tracking (Prompt Tracking)
Monitoring whether and how AI assistants mention or cite your brand by running a fixed set of prompts on a schedule and recording the answers.
AI Search & Agents
AI Referral Traffic
Website visits that arrive from AI assistants and AI search tools, such as ChatGPT or Perplexity, and are identifiable as such in analytics.
AI Search & Agents
AI Share of Voice
The share of AI answers, mentions or citations your brand earns across a set of prompts, compared with competitors.
AI Search & Agents
AI Visibility (LLM Visibility)
How often, and how well, a brand or page is retrieved, mentioned and cited in answers from AI search systems and LLM assistants.
Marketing Automation
AI Workflow Automation
Using machine learning models, APIs and scripts to take repetitive steps out of a marketing or SEO workflow, with people reviewing the output.
Core Concepts (AI/ML)
Algorithm
A defined set of steps for solving a problem; in machine learning, the procedure that learns patterns from data to produce a model.
APIs
Amazon Comprehend
AWS's pre-trained NLP API that extracts entities, key phrases, sentiment, language and personal data from text, with a confidence score for each resul…
Task Types
Anomaly Detection
Automatically spotting data points or periods that differ sharply from the normal pattern, such as an unexpected traffic drop or spike.
Search & SEO Concepts
Answer Accuracy Score (Search)
How often a model's predictions are correct; in Google's short-answer patent, a predicted score that decides whether a short answer is shown.
AI Search & Agents
Answer Engine Optimisation (AEO)
Answer engine optimisation (AEO) is making content easy for answer engines, from featured snippets to AI assistants, to retrieve, trust and quote as a…
APIs
API (Application Programming Interface)
A set of rules that lets one piece of software request data or functions from another, such as pulling Search Console data or calling an NLP model.
APIs
API Key
A secret code that identifies your project when a script or tool calls an API, so the provider can authorise the request and bill your account.
Tools & Platforms
Apps Script
Google's cloud-based JavaScript platform for automating Google Sheets and other Workspace apps, often used to call ML APIs from a spreadsheet without…
Core Concepts (AI/ML)
Artificial Intelligence (AI)
The broad field of building systems that appear to reason, learn and act intelligently; machine learning is one approach within it.
Generative AI & LLMs
Attention Mechanism
The technique that lets a neural network weigh how much each word in a text matters to every other word, so it reads words in context.
Search & SEO Concepts
Augmented Search Queries
Extra queries a search engine (or user) builds from the entities and attributes in an original query, used to expand or refine the results shown.
Core Concepts (AI/ML)
Bag of Words
A simple way to turn text into numbers by counting how often each word appears in a document, ignoring grammar and word order.
ML Models & Algorithms
BERT (Bidirectional Encoder Representations from Transformers)
Google's 2018 transformer language model that reads text in both directions to produce context-aware embeddings; the basis of many SEO NLP tools.
ML Models & Algorithms
BERTopic
A modular topic modelling technique that clusters transformer embeddings and labels each cluster with class-based TF-IDF to produce readable topics.
Core Concepts (AI/ML)
Bias (AI/ML)
When a model systematically favours some outcomes or groups over others because of skewed data, labels, design choices or how it is used.
Tools & Platforms
BigQuery
Google Cloud's serverless data warehouse for storing and querying very large datasets with SQL, including Search Console and GA4 exports.
Task Types
Binary Classification
A supervised machine learning task where the model sorts each item into one of exactly two classes, such as yes or no, spam or not spam.
ML Models & Algorithms
BIRCH (Balanced Iterative Reducing and Clustering using Hierarchies)
BIRCH (Balanced Iterative Reducing and Clustering using Hierarchies) clusters very large numeric datasets by summarising them in one pass into a tree.
ML Models & Algorithms
BM25
A classic ranking function that scores documents by how often query words appear, adjusted for word rarity and document length.
SEO & Marketing Metrics
Bounce Rate
In GA4, the percentage of sessions that were not engaged: under 10 seconds, no key event and only one page or screen view. It is the inverse of engage…
ML Models & Algorithms
Boyer-Moore
A fast exact string-search algorithm that compares a pattern from right to left and skips ahead after mismatches.
Entities & Knowledge Graphs
Brand Entity
A brand entity is how search engines and AI systems represent your organisation as a distinct, recognised thing, with its attributes and relationships…
ML Models & Algorithms
c-TF-IDF
Class-based TF-IDF: treats each cluster as one document to find the words that best describe it, used by BERTopic to name topics.
ML Models & Algorithms
Centroid-based Clustering
A clustering family that groups data around central points (centroids), each the average of its cluster's members. K-means is the best-known example.
Generative AI & LLMs
ChatGPT (OpenAI GPT models)
OpenAI's chatbot and its GPT model family: strong at summarising and rewriting, unreliable for structured analysis like entity extraction at scale.
AI Search & Agents
ChatGPT Search
ChatGPT's built-in web search, which retrieves current pages via OAI-SearchBot and answers with cited source links.
AI Search & Agents
Chunking (Content Chunking)
Splitting a page into smaller passages that retrieval systems index and score on their own; in AI search, the chunk is what gets retrieved and cited.
Marketing Automation
Churn Prediction
Using a model to estimate which customers or users are likely to stop buying, cancel or go inactive, so you can act before they leave.
Tools & Platforms
Claude (Anthropic)
Anthropic's family of large language models and AI assistant, used for writing, analysis, coding and agentic work, with web search and an API.
Task Types
Clustering (ML Task)
An unsupervised ML task that groups unlabelled data, such as keywords, pages or customers, by similarity so patterns emerge without predefined categor…
AI Search & Agents
Co-citation
When two sources or brands are cited or mentioned together by a third source, signalling they belong to the same topic or neighbourhood.
Model Evaluation
Coherence Score
A metric for how interpretable a topic model's topics are, based on how often each topic's top words appear together in real text.
Task Types
Computer Vision (Image Recognition)
The field of AI that lets machines interpret images and video, from recognising objects and text to describing scenes and finding similar images.
Generative AI & LLMs
Confidence Score (Generative AI)
A number, usually 0 to 1, showing how certain a model is about a label or prediction; low scores flag outputs to check before using them.
NLP (Concepts & Pipeline)
Content Moderation
Using NLP to flag text that may be harmful or sensitive, such as toxic, insulting or violent content, so it can be reviewed, hidden or reported.
Task Types
Content Transformation (ML)
Using ML and generative AI to convert content between formats or adapt it for new platforms, e.g. video to text, or a post to a script.
Generative AI & LLMs
Context Window
The maximum amount of text, measured in tokens, that a language model can consider at once, including the prompt, retrieved content and its answer.
ML Models & Algorithms
Convolutional Neural Network (CNN)
A deep learning architecture that scans data with small learned filters to detect local patterns; the classic model for image recognition.
Model Evaluation
Cosine Similarity
A measure of how closely two vectors point in the same direction, used to score how similar two texts, keywords or images are once they are embedded.
SEO & Marketing Metrics
CPC (Cost Per Click)
The price an advertiser pays for one click on a paid ad; in keyword research, an estimate SEOs use as a proxy for a query's commercial value.
SEO & Marketing Metrics
CTR (Click-Through Rate)
The share of impressions that result in a click (clicks divided by impressions). In SEO it shows how often searchers choose your result when they see…
Marketing Automation
Customer Segmentation (ML)
Grouping customers by shared behaviour or traits with machine learning clustering, so each group can get its own content, offers and messages.
AI Search & Agents
Dark AI Traffic
Visits sent or influenced by AI assistants that arrive without a referrer, so analytics records them as direct traffic instead of AI.
Core Concepts (AI/ML)
Data Labelling (Annotation)
Adding the correct answer (a category, entity, sentiment or score) to examples, so a supervised model can learn from them or be checked against them.
Core Concepts (AI/ML)
Data Preprocessing (Cleaning)
Cleaning, fixing and reshaping raw data so a machine learning model or analysis can use it, from removing duplicates to tokenising text and scaling nu…
APIs
DataForSEO API
A pay-as-you-go suite of SEO data APIs for SERPs, keywords, backlinks, on-page data and AI search mentions, used to automate research.
ML Models & Algorithms
DBSCAN
Density-based clustering that groups closely packed points, finds the number of clusters itself and labels isolated points as noise.
ML Models & Algorithms
Decision Tree
A supervised model that predicts by asking a series of yes/no questions about the data, branching like a flowchart until it reaches an answer.
Core Concepts (AI/ML)
Deep Learning
A branch of machine learning that trains many-layered neural networks on large datasets; it powers modern language models, image recognition and gener…
AI Search & Agents
Deep Research
An AI agent mode that runs many searches over several minutes, reads and reasons across sources, and returns a long, cited report.
Generative AI & LLMs
DeepSeek (R1 and V-series)
DeepSeek's open-weight reasoning LLM, released January 2025. MLforSEO benchmarked it for entity extraction: generous output, but inferred entities.
ML Models & Algorithms
Density-based Clustering
Clustering that groups points in dense regions, finds the number of clusters itself and leaves isolated points unassigned as noise.
Core Concepts (AI/ML)
Dimensionality Reduction
Compressing data with many features, such as embeddings, into fewer dimensions while keeping the important structure, for clustering or visualisation.
ML Models & Algorithms
Distance-based matching
Fuzzy string matching that scores how many character edits separate two strings, so near matches with typos or small spelling changes can be found.
ML Models & Algorithms
DistilBERT
A smaller, faster distilled version of BERT; in SEO it powers Mark Williams-Cook's classifier that predicts Google's eight query semantic classes.
ML Models & Algorithms
Distribution-based Clustering
A clustering family that assumes data comes from a mix of probability distributions, such as Gaussians, and assigns each point a probability for each…
Search & SEO Concepts
E-E-A-T
Experience, Expertise, Authoritativeness and Trust: the qualities Google's quality raters assess in content and creators, with trust at the centre.
Tools & Platforms
Elasticsearch
An open-source search and analytics engine for indexing large amounts of text, with fuzzy, phonetic and vector search built in.
Core Concepts (AI/ML)
Embedding
A list of numbers (a vector) that represents the meaning of a word, text, image or entity, so similar things sit close together.
ML Models & Algorithms
Embedding Model
An embedding model converts text, images or other content into vectors of numbers so that items with similar meaning end up close together.
NLP (Concepts & Pipeline)
Emotion detection/analysis
An NLP task that labels text with specific emotions such as joy, anger, sadness, fear and disgust, going beyond positive or negative sentiment.
Model Evaluation
Emotion Scores
Per-emotion scores (joy, sadness, fear, disgust, anger) an emotion analysis API assigns to a text, each from 0 to 1, showing how likely that emotion i…
ML Models & Algorithms
Encoder Model
A neural network that turns text or other input into numeric representations. In a Google patent, the encoder builds the knowledge used to answer ques…
Entities & Knowledge Graphs
Entity
A distinct, identifiable thing such as a person, place, organisation, product or concept, which search engines and LLMs use to understand meaning.
Entities & Knowledge Graphs
Entity Audit
A structured review of how search engines and AI systems recognise, connect and trust the entities on your site, with a prioritised list of fixes.
Entities & Knowledge Graphs
Entity Disambiguation
Working out which real-world entity a name refers to, such as Apple the company or apple the fruit, using context, schema and external identifiers.
NLP (Concepts & Pipeline)
Entity Extraction (NER)
Named entity recognition: finding names of people, organisations, places, products and other things in text and labelling each with a type.
Entities & Knowledge Graphs
Entity Hub
A page built to be the definitive source on one entity (a product, person, concept or organisation) instead of around one keyword.
Entities & Knowledge Graphs
Entity Linking (Reconciliation)
Entity linking connects a name mentioned in text or data to the exact entity it refers to in a knowledge base, such as a Wikidata item or Knowledge Gr…
Entities & Knowledge Graphs
Entity Relationships (Relationship Mapping)
The labelled connections between entities, such as a person who works for a company, that let search engines and AI systems see a connected graph.
Model Evaluation
Entity Salience Score
A 0 to 1 score from Google's Natural Language API showing how central each extracted entity is to the analysed text. Higher means more important.
NLP (Concepts & Pipeline)
Entity Sentiment Analysis
NLP that scores the sentiment expressed towards each entity in a text, so one review can be positive about one thing and negative about another.
Entities & Knowledge Graphs
Entity SEO
Optimising how search engines and LLMs recognise, understand and connect the entities behind your site: your brand, people, products and topics.
Entities & Knowledge Graphs
Entity Types
Entity types are the categories an entity belongs to, such as person, organisation, location, product or event, used by NLP tools, schema and knowledg…
Entities & Knowledge Graphs
Entity-Attribute-Variable (EAV) Model
A framework that breaks topics into entities, their attributes and the variables (values) those attributes take, e.g. dog, breed, Labrador.
Core Concepts (AI/ML)
Explainable AI (Explainability)
Explainable AI is the ability to understand and explain why a model produced a particular output, through interpretable models or explanation methods.
Model Evaluation
F1 Score
A single evaluation metric that balances precision and recall (their harmonic mean); useful when classes are imbalanced.
Core Concepts (AI/ML)
Feature (Machine Learning)
An input variable a model uses to make predictions, such as a page's word count, a keyword's search volume or a customer's order frequency.
Core Concepts (AI/ML)
Feature Extraction
Turning raw data such as text or images into numerical features, like word weights or embeddings, that a machine learning model can work with.
Search & SEO Concepts
Featured Snippet
A featured snippet is a highlighted answer Google shows near the top of the results, quoting a passage from one web page in reply to a question.
Learning Paradigms
Few-shot Prompting
Showing a language model a handful of worked examples inside the prompt so it copies the pattern, format or labels you want, without retraining.
Learning Paradigms
Fine-tuning
Training an existing pre-trained model a little further on your own labelled examples so it performs better on your specific task or domain.
Generative AI & LLMs
Function Calling (Tool Use)
Function calling lets a language model ask your code to run a defined tool, such as an API or database query, and use the result in its answer.
ML Models & Algorithms
Fuzzy Matching / Fuzzy String Matching
Scoring how similar two strings are when they are not identical, to catch typos and variants; it compares characters, not meaning.
Tools & Platforms
FuzzyWuzzy/RapidFuzz/PolyFuzz
Python libraries for fuzzy string matching. RapidFuzz is the fast default, TheFuzz is FuzzyWuzzy's successor, and PolyFuzz compares several matching m…
ML Models & Algorithms
Gaussian Mixture Models (GMM)
A soft clustering model that treats data as a mix of bell-shaped (Gaussian) distributions and gives each point a probability for every cluster.
Generative AI & LLMs
Gemini (Google)
Google's family of multimodal large language models, behind the Gemini app, the Gemini API and AI features in Google Search.
Generative AI & LLMs
Generative AI
AI models that create new content (text, images, audio, code) by predicting likely outputs from patterns learned in their training data.
AI Search & Agents
Generative Engine Optimisation (GEO)
Optimising content and brand signals so generative AI search systems retrieve, use and cite you in their answers.
Tools & Platforms
Gensim
An open-source Python library for topic modelling and document similarity, best known for LDA, LSI, word2vec and phrase detection on large text sets.
APIs
Google Autocomplete APIs
Google endpoints that return autocomplete predictions for Search, YouTube and Places queries, used to collect real query patterns for keyword research…
Tools & Platforms
Google Cloud
Google's cloud computing platform. For SEOs it hosts the ML APIs used most often: Natural Language, Knowledge Graph Search, Vision and Gemini models.
Tools & Platforms
Google Cloud AutoML
Google's tools for training custom models without writing model code; AutoML for text was retired in Vertex AI in favour of tuning Gemini.
APIs
Google Cloud Natural Language API (Google NLP API)
Google Cloud's pre-trained NLP API for entity analysis, sentiment, syntax, text classification and content moderation, usable without training a model…
Tools & Platforms
Google Colab
Google's free, browser-based Python notebook that runs code on Google's servers, so you can use ML scripts without installing anything.
APIs
Google Knowledge Graph Search API
Google's read-only API for looking up entities in its Knowledge Graph, returning names, types, descriptions and Knowledge Graph IDs.
APIs
Google Search Console API
Google's API for pulling Search Console performance data (clicks, impressions, CTR, position) and URL index status into your own tools.
SEO & Marketing Metrics
Google Trends
Data showing how interest in a search term or topic changes over time and by place, mainly from Google Trends, used to spot seasonality and rising top…
AI Search & Agents
Grounding
Anchoring a generative AI answer in verifiable sources supplied at request time, such as search results or a knowledge graph, instead of memory alone.
ML Models & Algorithms
Hard Clustering
A clustering approach where every data point belongs to exactly one cluster, producing non-overlapping groups. K-means is the best-known example.
ML Models & Algorithms
HDBSCAN (Hierarchical Density-Based Spatial Clustering of Applications with Noise)
A density-based clustering algorithm that finds clusters of varying density by itself and labels points that fit no cluster as noise.
ML Models & Algorithms
Hierarchical Clustering
Clustering that builds a tree of nested groups by repeatedly merging or splitting clusters, shown as a dendrogram you can cut at any level.
Tools & Platforms
Hugging Face Transformers
Open-source Python library for downloading and running pre-trained transformer models from the Hugging Face Hub for text, image and audio tasks.
Learning Paradigms
Human-in-the-loop
An approach where people stay involved in an AI or ML system, labelling data, reviewing outputs or approving actions, so results stay accurate.
Core Concepts (AI/ML)
Hyperparameter Tuning
Adjusting the settings you choose before training a model, such as the number of topics or clusters, to get better, more useful results.
APIs
IBM Watson NLU (Natural Language Understanding)
IBM's NLP API for entities, sentiment, emotion, relations, keywords and categories; deprecated in July 2026 and due to be removed in January 2028.
Search & SEO Concepts
Implicit User Feedback
Implicit user feedback is search behaviour, such as clicks, dwell time and query reformulation, that search systems read as signals of satisfaction.
APIs
IndexNow
An open protocol that lets sites instantly notify Bing, Yandex and other participating engines when URLs are added, updated or deleted.
Core Concepts (AI/ML)
Inference
Using a trained model to produce an output, such as a prediction, classification or generated answer, for new input it has not seen before.
Search & SEO Concepts
Information Gain (IG)
How much new, useful information something adds beyond what is already known; in SEO, a page's novelty compared with pages a user has already seen.
AI Search & Agents
Information Retrieval
The field and process of finding the documents or passages that best satisfy an information need in a large collection, the core of every search engin…
Search & SEO Concepts
Internal Linking
Internal linking is linking between pages on the same website, which helps people and search systems find content, understand topics and judge importa…
ML Models & Algorithms
Jaccard Similarity
A similarity measure for two sets: the number of shared items divided by the total number of distinct items, from 0 (nothing shared) to 1 (identical).
Search & SEO Concepts
JSON-LD
JSON for Linking Data: the script-based format Google recommends for structured data, describing a page's entities in a block separate from the HTML.
ML Models & Algorithms
K-Means Clustering
An unsupervised algorithm that splits data into a chosen number (k) of clusters by assigning each point to its nearest centre. Simple, fast and widely…
ML Models & Algorithms
KeyBERT
A keyword extraction library that uses BERT embeddings to find the words or phrases most similar in meaning to a whole text.
Search & SEO Concepts
Keyword Clustering
Grouping keywords that share a meaning, intent or set of ranking URLs, so one page or section can target each group.
SEO & Marketing Metrics
Keyword Difficulty (KD)
A tool-specific 0 to 100 estimate of how hard it would be to rank in Google's top 10 for a keyword, based mostly on the links of current top pages.
Task Types
Keyword Extraction
An NLP task that automatically pulls the most relevant words or phrases out of a text, used to summarise, tag and cluster content and keywords.
Search & SEO Concepts
Keyword Mapping
Assigning each keyword to the topic, cluster or page that should serve it, so your keyword research has a clear home and pages do not compete.
Search & SEO Concepts
Keyword Universe
The complete, labelled database of queries a brand cares about, organised by intent, entities and topics so teams can analyse and act on it.
Generative AI & LLMs
Knowledge Cutoff
The date after which a language model has no information from its training data, so newer facts must come from search or retrieval.
Entities & Knowledge Graphs
Knowledge Graph
A network of entities and the relationships between them; Google's Knowledge Graph powers knowledge panels, direct answers and entity understanding.
Entities & Knowledge Graphs
Knowledge Graph ID (KGMID)
The unique machine ID Google assigns to an entity in its Knowledge Graph, such as /m/0dl567 or /g/11…, used to identify it unambiguously.
Entities & Knowledge Graphs
Knowledge Panel
A Google results box summarising facts about a recognised entity, generated automatically from Google's Knowledge Graph.
Generative AI & LLMs
Large Language Model (LLM)
A neural network trained on huge amounts of text to predict the next token, which lets it generate, summarise, classify and answer questions.
ML Models & Algorithms
LDA (Latent Dirichlet Allocation)
A classic probabilistic topic model that treats each document as a mix of topics and each topic as a mix of words; you set the number of topics.
NLP (Concepts & Pipeline)
Lemmatisation and Stemming
Two ways to reduce words to a base form so variants count as one: stemming chops off endings, lemmatisation returns the dictionary form.
ML Models & Algorithms
Levenshtein Distance
The minimum number of single-character insertions, deletions or substitutions needed to turn one string into another.
NLP (Concepts & Pipeline)
Lexical or morphological analysis
The first phase of NLP: splitting text into words and analysing their structure (roots, prefixes, suffixes) so variants like 'shoes' and 'shoe' line u…
ML Models & Algorithms
LinkBERT
A BERT-style language model pretrained with hyperlinks between documents. SEOs use its embeddings to find semantically related pages for internal link…
Model Evaluation
LLM-as-a-Judge
Using a large language model to score or compare outputs against a rubric, as a scalable stand-in for human review.
AI Search & Agents
llms.txt
A proposed Markdown file at a site's root that lists its key pages for LLMs and AI agents; voluntary, with little evidence AI search uses it.
ML Models & Algorithms
Logistic Regression
A simple, interpretable classification model that turns a weighted sum of features into a probability between 0 and 1.
Tools & Platforms
Looker Studio
Google's free dashboard and reporting tool, renamed back to Data Studio in 2026, used to visualise SEO and ML analysis outputs.
ML Models & Algorithms
LSA (Latent Semantic Analysis)
A topic modelling technique that uses singular value decomposition (SVD) on a term-document matrix to uncover hidden themes in a set of documents.
Core Concepts (AI/ML)
Machine Learning (ML)
A branch of AI in which systems learn patterns from data to make predictions or decisions, instead of following hand-written rules.
Entities & Knowledge Graphs
Macro-Context
The broad domain a query or document belongs to, such as health or sport, which search systems use to narrow relevance before finer matching.
ML Models & Algorithms
Maximal Marginal Relevance (MMR)
A re-ranking method that picks results that are relevant to a query but different from each other, reducing redundancy in keywords, topics or search r…
ML Models & Algorithms
Mean Shift
A density-based clustering algorithm that moves points towards the densest nearby region, finding the number of clusters automatically.
ML Models & Algorithms
Metaphone
A phonetic algorithm that encodes words by how they sound in English, so differently spelled names that sound alike can be matched.
Search & SEO Concepts
Micro Intent
The specific need inside a broad search intent, such as budget, urgency or an exact product spec; often brand-specific and missed by generic intent la…
Entities & Knowledge Graphs
Micro-Context
The fine-grained layer of context: the specific terms, entities and subtopics that pin down exactly what a query or page is about within its broader d…
Core Concepts (AI/ML)
Model (AI/ML Model)
A machine learning model is the trained output of an algorithm: a set of learned patterns that turns new input data into a prediction, label, score or…
AI Search & Agents
Model Context Protocol (MCP)
An open standard for connecting AI applications to external tools and data, so an assistant can call an API, query a database or read files.
Task Types
Multi-Class Classification
Classification where each item gets exactly one label from three or more options, such as tagging a page as blog, product, FAQ or other.
Task Types
Multi-Label Classification
Classification where one item can receive several labels at once, such as tagging a post as both "analytics" and "SEO".
AI Search & Agents
Multimodal Search
Multimodal search lets people search with more than one type of input, such as text plus images, voice or video, and get answers drawn from all of the…
ML Models & Algorithms
MUM (Multitask Unified Model)
A multimodal, multilingual AI model Google announced in 2021 that can understand and generate language; used in Search for specific applications.
NLP (Concepts & Pipeline)
N-gram (Unigram, Bigram, Trigram)
A contiguous sequence of n words or characters from a text: unigrams (1), bigrams (2), trigrams (3) and so on, used to find recurring patterns.
ML Models & Algorithms
N-gram Matching
A fuzzy matching method that splits strings into overlapping chunks of n characters or words and scores similarity by how many chunks they share.
Tools & Platforms
n8n
A workflow automation platform that links apps, APIs and AI models into visual node-based workflows, run in the cloud or self-hosted.
NLP (Concepts & Pipeline)
Natural Language Processing (NLP)
The field of AI that enables computers to process, analyse and generate human language, from tokenisation and parsing to entities, sentiment and LLMs.
NLP (Concepts & Pipeline)
Natural Language Understanding (NLU)
The part of NLP focused on working out what text means: its intent, entities, sentiment and relationships, rather than only processing its words.
ML Models & Algorithms
Neural Network
A machine learning model made of layers of connected nodes that learn weights from data; the basis of deep learning, BERT and LLMs.
Tools & Platforms
NLTK (Natural Language Toolkit)
A free, open-source Python library for natural language processing tasks such as tokenisation, stop word removal, stemming and tagging.
ML Models & Algorithms
NMF (Non-negative Matrix Factorization)
A topic modelling method that splits a document-word matrix into document-topic and topic-word weights, all non-negative, for readable topics.
Tools & Platforms
No-code / Low-code ML
No-code and low-code ML means using machine learning through templates, spreadsheets, web apps and visual tools instead of writing (much) code yoursel…
Entities & Knowledge Graphs
Ontology
A formal specification of the concepts in a domain and the relationships between them; the blueprint a knowledge graph fills with real data.
Generative AI & LLMs
Open-source (Open-weight) LLM
A large language model whose trained weights are published for anyone to download, run and adapt, often under an open licence.
APIs
OpenAI API
OpenAI's paid developer interface for using GPT, embedding, image and speech models from your own code, billed per token or per use.
Entities & Knowledge Graphs
Orphan Entity
An entity your site mentions repeatedly but never defines on a dedicated page, leaving search engines and AI systems nothing authoritative to cite.
Model Evaluation
Overfitting
Overfitting is when a model learns its training data too closely, including noise, so it scores well in training but performs poorly on new data.
ML Models & Algorithms
PageRank
PageRank is the link analysis algorithm behind early Google that scores a page's importance from the number and importance of the pages linking to it.
Core Concepts (AI/ML)
Parameters and Weights
The internal numbers a model learns during training; weights set how strongly each input influences an output, and their count describes a model's siz…
NLP (Concepts & Pipeline)
Part-of-Speech (POS) Tagging
Part-of-speech tagging labels each word in a text with its grammatical role, such as noun, verb or adjective, as a step in analysing language.
AI Search & Agents
Passage Ranking
Scoring individual sections of a page for relevance, so a specific passage deep in a long page can surface for a narrow query.
ML Models & Algorithms
PCA (Principal Component Analysis)
A dimensionality reduction technique that compresses many features into a few components capturing the most variance, often to plot clusters in 2D.
Search & SEO Concepts
People Also Ask (PAA)
Google's SERP feature showing a group of related questions; expanding one reveals a short answer from a web page and often loads more questions.
Tools & Platforms
Perplexity
An AI answer engine that searches the web in real time and returns a synthesised answer with numbered citations to its sources.
ML Models & Algorithms
Phonetic matching
Fuzzy matching that compares how words sound, using codes like Soundex or Metaphone, so differently spelled names that sound alike match.
APIs
Place API
Google Maps Platform's Places API, whose Place Autocomplete and Query Autocomplete return location-aware suggestions you can mine for keywords.
Learning Paradigms
Pre-training
The first, large-scale training stage in which a model learns general patterns from huge datasets before it is adapted or used for specific tasks.
Model Evaluation
Precision and Recall
Precision is the share of a model's positive results that are correct; recall is the share of all true positives it finds.
Marketing Automation
Predictive Lead Scoring
Using a model trained on past conversions to estimate how likely each new lead is to become a customer, so sales can prioritise.
ML Models & Algorithms
Predictive Text Models
Models that predict the rest of a word or query as someone types, such as Google and YouTube autocomplete, and a rich source of real query ideas.
Search & SEO Concepts
Programmatic SEO
Programmatic SEO is creating many search-targeted pages from a template and a structured dataset, so each page covers one entity or attribute combinat…
ML Models & Algorithms
Programmer Model
In a Google patent, the model that turns a natural-language question into a short program, which is run against a knowledge graph to find the answer.
Generative AI & LLMs
Prompt Engineering
Prompt engineering is designing and testing the instructions, context and examples you give a language model so it returns reliable, usable output.
Tools & Platforms
Python
A free, readable programming language and the default choice for data analysis, machine learning and SEO automation scripts.
APIs
PyTrends (Google Trends API)
An unofficial Python library that pulls Google Trends data, such as interest over time, by region and related queries, into your own code.
Search & SEO Concepts
Query Augmentation
How search engines expand or refine a query with related entities and attributes to return richer results. It is the conceptual ancestor of AI query f…
Search & SEO Concepts
Query Context
The meaning and scope a user gives a single search through its words, modifiers and hints, plus signals like location and device.
Model Evaluation
Query Distance
How far apart two search queries are, measured by spelling (fuzzy matching), by meaning (embeddings) or by the results they share.
AI Search & Agents
Query Fan-Out
The process where an AI search system expands one query into many related sub-queries, retrieves results for each, and combines them into one answer.
Search & SEO Concepts
Query Path
The full journey of searches a person makes to reach a goal, often across several sessions, moving from broad to specific queries or sideways.
Search & SEO Concepts
Query Sequence
The order of related searches a person makes within one session. Search engines use these sequences to infer intent and suggest the next search.
Generative AI & LLMs
Qwen (Alibaba)
An Alibaba Cloud large language model from the Qwen 2.5 generation, offered through Qwen Chat and API, and tested by MLforSEO for entity extraction.
ML Models & Algorithms
Random Forest
A supervised algorithm that trains many decision trees on random samples of the data and combines their votes, for more accurate, stable predictions.
Entities & Knowledge Graphs
RDF (Resource Description Framework)
A W3C data model that describes information as subject-predicate-object triples, the foundation under JSON-LD, schema.org and Wikidata.
AI Search & Agents
Re-ranking
A second, more precise scoring pass that reorders a shortlist of retrieved results before they are shown or passed to an AI model.
Generative AI & LLMs
Reasoning Model
A reasoning model is a language model trained to work through a problem step by step, spending extra computation on "thinking" before it gives a final…
Search & SEO Concepts
Redirect Mapping
Matching every old or broken URL to its best live equivalent and setting redirects, so users and search engines reach the right page after changes.
Task Types
Regression (ML Task)
A supervised machine learning task that predicts a continuous number, such as traffic, revenue or click-through rate, from input features.
Tools & Platforms
Regular Expressions (Regex)
A compact pattern language for finding, matching and extracting text, used in Search Console filters, Google Sheets, GA4 and Python.
Learning Paradigms
Reinforcement Learning
A machine learning approach where an agent learns by trial and error, taking actions and adjusting its behaviour based on rewards or penalties.
AI Search & Agents
Retrieval-Augmented Generation (RAG)
A method where an AI system retrieves relevant passages from an index first, then writes its answer from them, so responses are grounded in sources.
Search & SEO Concepts
Rich Results
Search results with extras beyond a plain link, such as review stars, prices, carousels and event dates, usually generated from structured data on the…
Search & SEO Concepts
Rich Snippets
Search results showing extra details from a page, such as review stars, price, availability or cooking time, usually generated from structured data.
Search & SEO Concepts
robots.txt
A text file at the root of a site that tells crawlers, including search and AI bots, which URLs they may and may not fetch.
Entities & Knowledge Graphs
sameAs (Schema Property)
sameAs is a Schema.org property that links an entity on your site to its profiles elsewhere (Wikidata, LinkedIn, Crunchbase) to confirm they are the s…
Search & SEO Concepts
Schema Markup (Schema.org)
Code that labels page content with the shared Schema.org vocabulary so search engines and AI systems can read what a page is about without guessing.
Tools & Platforms
scikit-learn
A free Python library for classic machine learning: clustering, classification, regression, preprocessing and model evaluation.
Tools & Platforms
Screaming Frog SEO Spider
Screaming Frog SEO Spider is a desktop website crawler for technical SEO audits that can also extract page content and run embeddings and AI prompts.
Search & SEO Concepts
Search Intent
The goal behind a search query, such as learning, finding a site, comparing options or buying, which shapes what content can satisfy it.
Task Types
Search Intent Classification
Labelling queries by what the searcher wants, such as informational, navigational, commercial or transactional, using rules, SERP signals or ML models…
Search & SEO Concepts
Search Personalisation
Tailoring search results or AI answers to a person's context, such as location, language, recent searches and past behaviour.
SEO & Marketing Metrics
Search Volume
The estimated average number of times a keyword is searched per month, used to size demand and prioritise keywords.
NLP (Concepts & Pipeline)
Semantic Analysis
The part of NLP concerned with meaning: what words and sentences mean and how they relate, covering tasks like entity analysis, sentiment and topics.
Search & SEO Concepts
Semantic Keyword Research
Semantic keyword research studies the meaning, entities, intent and journeys behind searches, instead of treating keywords as isolated strings with vo…
Core Concepts (AI/ML)
Semantic Representation
A numerical form of text that captures its meaning so machines can compare it, such as a bag of words, a TF-IDF vector or an embedding.
AI Search & Agents
Semantic Search
Search that matches the meaning and intent behind a query, using entities, context and embeddings, rather than only the exact words typed.
Search & SEO Concepts
Semantic SEO
Optimising for meaning rather than exact keywords: covering entities, their attributes, relationships and the intent behind searches.
Model Evaluation
Semantic Similarity
A measure of how close two texts are in meaning, usually the cosine similarity of their embeddings, whether or not they share words.
Entities & Knowledge Graphs
Semantic Triple (Tuple)
A single fact in a knowledge graph, written as subject, predicate and object (for example Paris, capital of, France), often with context such as sourc…
Tools & Platforms
Sentence Transformers (library)
Open-source Python library (also called SBERT) for turning text into embeddings and scoring similarity; it powers many MLforSEO notebooks.
ML Models & Algorithms
Sentence-BERT (sBERT)
A BERT-based model that turns sentences or short phrases into embeddings that can be compared with cosine similarity, ideal for mapping keywords to to…
NLP (Concepts & Pipeline)
Sentiment Analysis
NLP that detects whether text expresses a positive, negative or neutral opinion, often with a score for polarity and strength.
Model Evaluation
Sentiment Magnitude
How strong the emotion in a text is, regardless of whether it is positive or negative; returned by Google's Natural Language API with the sentiment sc…
Model Evaluation
Sentiment Score (Polarity)
A number showing whether text is positive, negative or neutral. In Google's Natural Language API it runs from -1.0 (very negative) to +1.0 (very posit…
Search & SEO Concepts
SERP Features
Any search result element beyond the ten blue links, such as featured snippets, People Also Ask, AI Overviews, videos, images and local packs.
APIs
Serpapi
A paid API (with a free tier) that returns search engine results pages as structured JSON, including People Also Ask and AI Overviews.
Search & SEO Concepts
Session Context
The signals a search engine gathers across a user's series of searches and interactions, such as earlier queries and clicks, to interpret the next one…
SEO & Marketing Metrics
Session Duration
How long a visit to your site lasts. In GA4, average session duration is the mean length, in seconds, of users' sessions.
Model Evaluation
Silhouette Score
A clustering quality metric from -1 to 1 that compares how close each point is to its own cluster versus the nearest other cluster.
Model Evaluation
Similarity Score
A number, usually 0 to 1 or 0 to 100, expressing how alike two strings, texts or vectors are; how it is calculated decides what 'similar' means.
ML Models & Algorithms
Soft/Fuzzy Clustering
A clustering approach where each data point gets a probability or degree of belonging to several clusters, instead of a single fixed group.
Tools & Platforms
spaCy
A free, open-source Python library for production NLP: tokenisation, part-of-speech tagging, dependency parsing, entity recognition and custom models.
NLP (Concepts & Pipeline)
Stop Words
Very common words such as "the", "and" and "of" that are often removed before text analysis so the meaningful terms stand out.
Search & SEO Concepts
Structured Data
Machine-readable markup, usually schema.org in JSON-LD, that tells search engines and AI systems which entities a page describes and how they relate.
Learning Paradigms
Supervised Learning
Machine learning trained on labelled examples, where the right answer is known, so the model can predict labels (classification) or numbers (regressio…
ML Models & Algorithms
Support Vector Machine (SVM)
A classification model that finds the boundary separating classes with the widest possible margin; strong on text features like TF-IDF.
NLP (Concepts & Pipeline)
Syntax analysis (parsing)
The NLP step that analyses a sentence's grammatical structure: parts of speech, word roots and how words depend on each other.
AI Search & Agents
Synthetic Queries
Synthetic queries are searches generated by a system, not typed by a user, used by AI search to expand, interpret and explore a question before answer…
Generative AI & LLMs
Temperature (LLM)
A generation setting that controls how random an LLM's word choices are: lower is more predictable, higher is more varied.
Task Types
Text Classification
A machine learning task that sorts text such as pages, keywords or reviews into predefined categories, learned from labelled examples.
Task Types
Text Summarisation
An NLP task that compresses long text into a shorter version that keeps the key points, either by extracting sentences or writing new ones.
ML Models & Algorithms
TF-IDF (Term Frequency-Inverse Document Frequency)
A weighting that scores how important a word is to one document compared with a whole collection, so distinctive words count more than common ones.
ML Models & Algorithms
TF-IDF with Cosine Similarity
Comparing texts by turning each into a TF-IDF vector, where rarer terms weigh more, and measuring the angle between vectors with cosine similarity.
Task Types
Time Series Forecasting
Predicting future values of a metric recorded over time, such as traffic, impressions or search demand, from its past patterns.
Generative AI & LLMs
Token (LLM)
A token is the small unit of text (a word, part of a word or a symbol) that a language model reads and writes; usage, limits and pricing are counted i…
NLP (Concepts & Pipeline)
Tokenization
Splitting text into smaller units called tokens, such as words, subwords or characters, so a model can process it. The first step in most NLP pipeline…
ML Models & Algorithms
Top2Vec
An embedding-based topic model that finds dense areas of document vectors and uses the nearest words to describe each topic.
Search & SEO Concepts
Topic Clusters
A content structure where a central pillar page covers a broad topic and links to, and from, focused pages on its subtopics.
NLP (Concepts & Pipeline)
Topic Modeling
An unsupervised technique that discovers the themes running through a large collection of text, such as pages, reviews or queries.
Search & SEO Concepts
Topical Authority
The perceived depth and trustworthiness of a site on a subject, built by covering a topic and its related entities thoroughly and consistently.
Model Evaluation
Train/Test Split & Cross-validation
Holding back data the model never trains on to test how well it generalises; cross-validation repeats this across several splits.
Core Concepts (AI/ML)
Training Data
The examples a machine learning model learns from; for LLMs, the large text corpus absorbed before release that shapes what the model knows.
Generative AI & LLMs
Transformer (Architecture)
The neural network architecture behind BERT, GPT, Gemini and most modern language models, built around attention instead of reading text word by word.
ML Models & Algorithms
UMAP
A dimensionality reduction algorithm that compresses high-dimensional data such as embeddings into a few dimensions while keeping similar items close.
AI Search & Agents
Universal Commerce Protocol (UCP)
An open standard led by Google for AI agents to discover products, build carts and check out with merchants; announced January 2026.
Learning Paradigms
Unsupervised Learning
Machine learning on data without labels, where the model finds groups, patterns or structure on its own, as in clustering and topic modelling.
Search & SEO Concepts
User Search Behavior
How people interact with search: what they type, reformulate, click, how long they stay and when they return; used by search engines as implicit feedb…
Tools & Platforms
Vector Database
A database that stores embeddings and finds the most similar ones fast, powering semantic search, recommendations and RAG.
AI Search & Agents
Vector Search
Finding content by meaning: queries and documents become embedding vectors, and the closest vectors are returned as the most similar results.
Tools & Platforms
Vertex AI
Google Cloud's machine learning and generative AI platform. In April 2026 it was folded into Gemini Enterprise Agent Platform.
AI Search & Agents
Visual Search (Google Lens)
Searching with an image or camera instead of words, as with Google Lens and Circle to Search; the image itself becomes the query.
Tools & Platforms
Web Scraping
Automatically extracting data from web pages, such as page copy, SERPs or reviews, into a structured dataset you can analyse or feed to a model.
AI Search & Agents
WebMCP
WebMCP is a proposed web standard that lets a website expose structured actions (tools) to AI agents in the browser, so they need not guess from the U…
APIs
Whisper (OpenAI)
OpenAI's speech-recognition model that transcribes and translates audio in many languages, available open source and through the OpenAI API.
Entities & Knowledge Graphs
Wikidata
A free, community-edited knowledge base of structured facts about entities, used by search engines and AI systems to recognise and verify who is who.
ML Models & Algorithms
Word2Vec
Word2Vec is a 2013 neural method from Google that learns a vector for each word from the contexts it appears in, so related words get similar vectors.
Tools & Platforms
Zapier
A no-code automation platform that connects thousands of web apps so that an event in one app triggers actions in others, including AI steps.
Search & SEO Concepts
Zero-click Search
A search that ends without the user clicking any result, because the answer was shown on the results page or in an AI response.
Learning Paradigms
Zero-shot Learning
Getting a model to perform a task or apply labels it was never trained on, with no examples, using only knowledge from pre-training.