clustering
when people or things are close together in groups
the forming, existence, or arrangement of people or things in close groups rather than spread evenly
coming together in a small, close group
forming a close group, or gathering close together around someone or something
a way for software to put similar data items into groups
a statistical or machine-learning process that groups data points according to similarity, often without using predefined labels
connecting computers so they can work together as one stronger or safer system
the practice of connecting servers or computers so that they work together as one system, often to improve performance or keep a service running if one machine fails
- Base: clustering
- Plural: clustering, clusterings
clustering algorithm
1data-grouping computer method
a computer method that puts similar data into groups
- We tested several clustering algorithms on the customer data.
- The clustering algorithm grouped similar news articles together.
- Choosing the right clustering algorithm depends on the shape of the data.
- The clustering algorithm found three groups.
- We used a clustering algorithm for the photos.
clustering coefficient
1network grouping measure
a number that shows how much a network forms close groups
- A high clustering coefficient means that a person's friends are likely to know one another.
- The network had a low clustering coefficient, so connections were widely spread out.
- Researchers compared the clustering coefficient of different social networks.
- The clustering coefficient is a number.
- A high clustering coefficient means many close connections.
clustering illusion
1seeing patterns in randomness
the mistake of thinking random groups are meaningful patterns
- After three wins in a row, he believed the slot machine was hot, a classic clustering illusion.
- The article warned investors about the clustering illusion in stock-market charts.
- People often see streaks in random data because of the clustering illusion.
- The clustering illusion makes random events look meaningful.
- A short winning streak can create a clustering illusion.
failover clustering
1backup-system service continuity
a setup where another computer takes over when one computer fails
- The bank uses failover clustering to keep its payment system online.
- Failover clustering allowed the database to recover within seconds.
- We tested failover clustering before moving the application to production.
- Failover clustering keeps the service running.
- If one server stops, failover clustering uses another server.
hierarchical clustering
1tree-like data grouping
a data-grouping method that shows groups inside larger groups
- The researchers used hierarchical clustering to compare gene-expression patterns.
- Hierarchical clustering produced a tree diagram of the documents.
- The chart from hierarchical clustering showed two large groups and several smaller ones.
- Hierarchical clustering showed small groups inside big groups.
- We used hierarchical clustering for the survey data.
k-means clustering
1centroid-based data grouping
a common data-grouping method where the user chooses how many groups to make
- K-means clustering divided the customers into five behavior groups.
- The analyst used k-means clustering because the dataset was large and numeric.
- K-means clustering works poorly when the clusters have very irregular shapes.
- K-means clustering made four groups.
- We chose k before using k-means clustering.
spatial clustering
1grouping in physical space
when things happen or exist close together in certain places
- The study found spatial clustering of asthma cases near major roads.
- Spatial clustering of tech firms has changed the local housing market.
- Police looked for spatial clustering in the pattern of burglaries.
- The map shows spatial clustering of restaurants.
- Spatial clustering means the cases are close together in place.
temporal clustering
1grouping close in time
when events happen close together in time
- The report noted temporal clustering of accidents during the winter months.
- Doctors investigated temporal clustering of the symptoms after the event.
- The data showed temporal clustering, with most complaints arriving on Mondays.
- The calls showed temporal clustering in the morning.
- Temporal clustering means events happen close together in time.