Expectation step

As explained in Data Without Labels

The expectation step of expectation-maximization estimates the probability that each observation belongs to each latent component using current model parameters.

The expectation step assigns each data point probabilistically to every Gaussian component in a Gaussian mixture model (GMM). It calculates the probability that each component generated the point, producing soft assignments rather than fixed cluster labels. section 5.6.1

Produces a probability matrix with:

• One row for each data point. section 5.6.1
• One column for each Gaussian component. section 5.6.1
• Normalized probabilities indicating how strongly each component explains each point. section 5.6.1

(5.10)

Inspect the probability formula underlying the normalized soft assignments. section 5.6.1

How Data Without Labels covers it

How deep it goes

Defines normalized probability calculations and their matrix output, then defers parameter updates to the maximization step.

Examples

Uses a dataset of n items and a data-point-by-Gaussian probability matrix.

What it assumes

Assumes Gaussian mixture components with means, covariances, and weights.

What you can do after

Evaluate how strongly each Gaussian component explains every observation.

How it connects

Connects GMM clustering with Gaussian distributions, soft assignments, and the maximization step.

liveBookmanning.com

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

Related concepts

• expectation maximizationpart of · Expectation maximization is an iterative algorithm for estimating model parameters with latent or unobserved variables by alternating latent-variable estimation with parameter optimization.
• soft assignmentproduces · Soft assignment gives an observation graded membership scores, probabilities, or weights across alternative classes or clusters.
• meanuses · A mean is a measure of central tendency, calculated as the arithmetic average of values or, for a random variable, as the probability-weighted average of its possible values.

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.