[go: up one dir, main page]
Scribd
Upload
1 views3 pages

5 A

The document compares hard assignment clustering (like K-means) with soft assignment clustering (model-based). K-means assigns data points to the nearest centroid and assumes spherical clusters, while model-based clustering uses statistical models to capture complex shapes and provides probabilistic cluster memberships. K-means is efficient for simple cases, whereas model-based clustering offers flexibility for overlapping and varied cluster shapes.

Uploaded by

SUBHASISH BEHERA
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
1 views3 pages

5 A

The document compares hard assignment clustering (like K-means) with soft assignment clustering (model-based). K-means assigns data points to the nearest centroid and assumes spherical clusters, while model-based clustering uses statistical models to capture complex shapes and provides probabilistic cluster memberships. K-means is efficient for simple cases, whereas model-based clustering offers flexibility for overlapping and varied cluster shapes.

Uploaded by

SUBHASISH BEHERA
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 3
Aspect Lec) Model-Based Clustering Clustering Clustering Hard assignment: Soft assignment: each Approach each data point eC Wolo) ae Lote fel oR elid probability of one cluster based belonging to each on closest centroid | cluster based ona (distance-based). statistical model. Assumptions | Assumes clusters Assumes data come are spherical and from a mixture of equally sized; no probability explicit data distributions (e.g., distribution Gaussian mixtures), assumed. PT am ered oie} estimated. Cluster Detects spherical Can detect clusters of ES] T ol) fol Slee na various shapes (elliptical, different covariance Sot.) B Output Partition of data Probabilistic cluster Taeme eee (ar Ue| labels). memberships and model parameters (means, covariances). Example Tey Lala<=tiCe lake d Palle Piss Cm a) of squares using centroid updates. Grouping customers into 3 clusters based on spending patterns using Euclidean distance. Maximum likelihood Clue cei Expectation- Maximization) based on mixture models. Clustering the Old Faithful geyser data into 3 clusters modeled as Gaussian distributions with different means and covariances. Ti) ol delat tel a) « K-means example: Suppose an online store groups customers into 3 clusters by spending frequency and amount. K-means assigns each customer to the nearest cluster centroid, iteratively updating centroids until stable (10). ¢ Model-based example: The Old Faithful geyser dataset can be modeled as a mixture of three Gaussian clusters with elliptical shapes, where each data point has a probability of belonging to each cluster. This approach captures more complex cluster shapes and provides probabilistic assignments (6). ST lag K-means is simple, fast, and effective for spherical clusters with hard assignments, while model-based clustering offers a flexible probabilistic framework that can model clusters with different shapes and overlapping boundaries, providing richer information about cluster it laales eal on

You might also like