Link prediction metric

As explained in Graph Algorithms for Data Science

A link prediction metric is a numerical score that estimates the likelihood of a prospective relationship between two graph nodes.

A link prediction metric is a numerical score for estimating whether two graph nodes may form a relationship. Common friends, Jaccard similarity, and other local-neighborhood overlap measures provide signals for prospective links. section 10.1

Use link prediction metrics in two practical ways:

• Rank node pairs by their scores and treat the highest-scoring pairs as likely future links. section 10.1
• Provide multiple metrics to a classification model so it can learn feature weights and useful combinations. section 10.1

Channel recommendations

For Twitch-style channel recommendations, shared audiences already indicate similarity; predicting future audience sharing can identify additional channel recommendations. section 10.1

Figure 10.4 Link prediction flow

Follow the link prediction flow from node-pair metrics toward predicted relationships and recommendation use. section 10.1

How Graph Algorithms for Data Science covers it

How deep it goes

Explains common-neighborhood metrics, score ranking, and classifier use without detailing metric implementations.

Examples

Uses common friends, Jaccard similarity, and future audience sharing between Twitch channels.

What you can do after

Apply scores to rank likely links or train a classifier for improved channel recommendations.

How it connects

Connects common friends and Jaccard similarity with neighborhood overlap, classification models, and recommendation systems.

liveBookmanning.com

Read Graph Algorithms for Data Science on liveBook Get the book at manning.com

Related concepts

• link predictionmeasures · Link prediction is a graph-learning task that estimates whether a connection exists or may occur between nodes, often by scoring candidate links.

This explainer was generated by Manning's knowledge graph from the text of Graph Algorithms for Data Science. The section links point to the parts of the book it draws on.