Preferential Attachment is a phenomenon in which **new nodes entering into a network attach themselves to other nodes preferentially based on some criteria** (i.e. not randomly). This is a naturally occurring phenomenon that happens in all sorts of real-world settings across any sort of a discipline which can be thought of using a [[Graph]]. Preferential attachment, even with very minor [[Bias]]es, naturally leads to the creation of **hubs - which are highly-connected nodes**, well beyond the average number of connections experienced by all nodes. > [!tldr] Hubs naturally occur as a network grows, as a result of *preferential attachment* Preferences for attachment may be directly based on number of existing attachments (i.e. "I want to find a node lots of others are connected to") or may be based on some other objective criteria --- but it will end up inevitably looking just like you're seeking nodes that lots of other nodes are connected to (e.g. "I'm betting on this horse because he's favored to win, and because he's favored to win most bets are on this horse"). This is another example of [[Snowball Effects]]. ## Examples Examples given in the source. - Chicago O'Hare airport - connect to hundreds of destinations - Atlantic Cod - connect hundreds of predators and prey - [[Adenosine Triphosphate|ATP]] - connected to hundreds of chemical reactions - The Prefrontal Cortex - connected to hundreds of functions Additional examples from my brain: - Good [[My Notes|notes]] are more likely to be referenced by other notes - There are 10 notes (as of 2025-10-09) that have 40+ backlinks, and 1800 with fewer than 10 ![[Hub_in_Notes.png]] **** # More ## Source -