K-means clustering

As explained in Data Without Labels

K-means clustering is an iterative clustering algorithm that partitions observations into a specified number of groups by minimizing within-group squared distance to cluster centroids.

K-means groups observations with similar numeric features into k nonoverlapping clusters. Each observation belongs to one cluster, identified by its nearest center, or centroid. It is useful when, for example, a retailer wants to find customer segments without supplying a target variable. section 2.3.1 section 2.10

The algorithm alternates between assignment and recalculation:

• Choose k, the number of clusters, and initialize k centers. section 2.3.1
• Assign every observation to its nearest center. section 2.3.1
• Move each center to the mean of its assigned observations, then reassign observations. Repeat until the clustering no longer improves. section 2.3.1

Figure 2.6 Step 1 represents the raw dataset. In step 2, the algorithm initiates three random centroids as we have given the input of the number of clusters as three. In step 3, all the neighboring points of the centroids are assigned to the same cluster.

The diagram makes the first assignment visible: three initial centroids produce three groups of nearby points. Later recalculations can change those assignments. section 2.3.1

Cluster vehicles in Python

Load the vehicle data into a pandas dataframe and check its shape, variable types, sample rows, class counts, and missing values. Exclude the class column from the clustering features. Standardize the remaining features with StandardScaler so a large-valued dimension does not dominate distance calculations. Plot average Euclidean distortion across candidate k values to look for an elbow; the illustrated fit uses KMeans(n_clusters=3, n_init=15, random_state=2345) on the standardized data. Inspect cluster_centers_, labels_, label counts, and a cluster-colored plot to understand the result. section 2.3.5

Treat the clusters as a starting point for interpretation, not automatic business categories. Choosing k can be difficult, random initial centers can change the output, and outliers can distort it; run multiple fits and examine the data before accepting a solution. Basic k-means also cannot use categorical variables directly. For customer segmentation, check cluster sizes and distinguishing attributes before naming groups or acting on them. section 2.3.4 section 2.6.4

How Data Without Labels covers it

How deep it goes

Works through centroid reassignment and a Python workflow; defers detailed outlier treatment.

Examples

Uses vehicle features for a Python fit and customer transactions for business segmentation; suggests an iris comparison exercise.

What it assumes

Introduces centroid assignment step by step while using Python dataframes, plotting, and feature scaling in the implementation.

What you can do after

Apply it via scikit-learn KMeans to standardized features, examine an elbow plot, and inspect centroids and cluster labels.

How it connects

Connects centroid-based clustering with distance metrics, feature standardization, the elbow method, hierarchical clustering, and DBSCAN.

liveBookmanning.com

Read Data Without Labels on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Advanced Algorithms and Data StructuresK-means is an iterative clustering algorithm that partitions observations into a specified number of clusters by minimizing within-cluster squared distances to cluster centroids.
• 

  liveBookmanning.com
  Statistics Slam DunkK-means is a clustering algorithm that partitions observations into a prespecified number of groups by minimizing within-group squared distances to cluster centroids.
• 

  liveBookmanning.com
  Classic Computer Science Problems in Swift: Essential techniques for practicing programmersK-means clustering is an iterative unsupervised learning algorithm that partitions data points into a predefined number of clusters by repeatedly assigning points to their nearest centroid and recomputing the centroids.
• 

  liveBookmanning.com
  The Well-Grounded Data AnalystK-means is a clustering algorithm that partitions observations into a specified number of groups by minimizing distances from observations to their assigned cluster centers.
• 

  liveBookmanning.com
  Build AI-Enhanced Web AppsK-means clustering is an unsupervised clustering algorithm that assigns data points to a chosen number of groups according to their distances from cluster centers.
• 

  liveBookmanning.com
  AI-Powered SearchK-means is an unsupervised clustering algorithm that partitions items into a specified number of groups by iteratively assigning items to and updating cluster centroids.

Compare how 7 books explain K-means clustering →

Related concepts

• centroiduses · A centroid is a central point representing a geometric figure or set of points, commonly calculated as the coordinate-wise mean, with weights when applicable.
• Euclidean distanceuses · Euclidean distance is the straight-line distance between two points in Euclidean space, calculated as the square root of the sum of squared differences between corresponding coordinates.
• clusteringproduces · Clustering is the technique of grouping data items according to similarity or shared characteristics.
• hierarchical clusteringcontrasts with · Hierarchical clustering is a clustering method that organizes observations into nested groups by successively merging or splitting clusters.
• unsupervised learningkind of · Unsupervised learning is a machine-learning approach that discovers patterns, structure, or representations in data without labeled target outputs.
• hard clusteringkind of · Hard clustering assigns each observation to at most one cluster, with all clusters represented at the same hierarchy level.

This explainer was generated by Manning's knowledge graph from the text of Data Without Labels. The section links point to the parts of the book it draws on.