{"id":34342,"date":"2022-05-24T11:16:47","date_gmt":"2022-05-24T11:16:47","guid":{"rendered":"https:\/\/www.arimetrics.com\/glosario-digital\/random-navigation-model"},"modified":"2026-09-30T11:50:58","modified_gmt":"2026-09-30T11:50:58","slug":"random-navigation-model","status":"publish","type":"encyclopedia","link":"https:\/\/www.arimetrics.com\/en\/digital-glossary\/random-navigation-model","title":{"rendered":"Random Navigation Model"},"content":{"rendered":"

\"RandomDefinition<\/strong>:<\/p>\n

The random navigation model<\/strong>, also known as the random surfer model, is a probabilistic interpretation of link navigation used to explain the basic operation of PageRank. It represents a hypothetical person who moves from one page to another by following links or jumping to another URL in the analysed set.<\/p>\n

It does not describe how a real user browses or predict a particular session. It is a mathematical model<\/strong> applied to a graph of pages and links to estimate the stable probability of being at each node after many steps.<\/p>\n\n

Relationship with PageRank<\/h2>\n

The random surfer model provides an intuitive way to interpret PageRank<\/a>. A page receives probability from pages linking to it and distributes its contribution<\/strong> among its outgoing links. Links are not counted as identical, isolated votes: their effect also depends on the probability accumulated by the source page.<\/p>\n

The result is a relative score<\/strong> for each page within the calculated graph. It is not the click probability of a person, expected traffic or the authority of an entire domain. Nor does it determine a position in a search engine<\/a> by itself.<\/p>\n

Probabilistic operation<\/h2>\n

The calculation can be described as a repeated sequence<\/strong> of decisions within the network:<\/p>\n

    \n
  1. Starting page:<\/strong> The process begins on a page in the analysed set.<\/li>\n
  2. Link following:<\/strong> At each step, one of the available outgoing links may be selected, usually with equal probability in the basic model.<\/li>\n
  3. Random jump:<\/strong> With the remaining probability, another page is selected even when there is no direct link from the current page.<\/li>\n
  4. Iteration:<\/strong> The process repeats until the probability distribution converges within the chosen criterion.<\/li>\n<\/ol>\n

    If a page has no outgoing links, the calculation needs a rule to prevent probability from remaining trapped. Depending on the formulation, these dangling nodes<\/strong> may be treated as linking to the set of pages or have their probability mass redistributed in an equivalent way.<\/p>\n

    Damping factor<\/h2>\n

    The damping factor<\/a> represents the probability of continuing to follow links. Its complement represents the random jump. This mechanism prevents the calculation from becoming trapped in cycles or closed areas and encourages a convergent distribution<\/strong>.<\/p>\n

    The classic explanation often uses 0.85 as an example factor, but this value is not part of the model’s universal definition. It is a calculation parameter<\/strong> that may be changed for a particular experiment or implementation.<\/p>\n

    The jump should not be interpreted literally as typing a URL into a browser. It is a mathematical device<\/strong> that ensures a possibility of reaching pages outside the link path currently being followed.<\/p>\n

    Graph interpretation<\/h2>\n

    The score depends on the complete structure being analysed. Its interpretation involves these graph elements<\/strong>:<\/p>\n