Instructions: This form is used to infer ontology axioms from variants for a given phrase and tag sequence.

• usage is_a _entity
• site_information is_a _information
• web_site_usage_information is_a web_site_information
• web_site_usage_information is_a site_usage_information
• usage_information same_as _information about usage
• web is_a _entity
• web_site_usage_information same_as _information about web_site_usage
• site_usage_information is_a usage_information
• usage_information is_a _information
• site_usage_information same_as _information about site_usage
• web_site_information is_a site_information
• site_usage is_a usage
• site is_a _entity
• site_information same_as _information about site
• web_site is_a site
• web_site_usage is_a site_usage

The tag sequence is encoded using the following rules:

1. Each word is assigned a letter codes as follows: T = thing or entity, A = agent, M = modifier, E = event or action, P = property, X = information, or _ (underscore) = the word is part of a compound word that inherits the last word’s code.
2. Normal processing treats adjacent words as hypernym-hyponym pairs, such that W₁W₂ yields the interpretation: W₁W₂W₃ is_a W₂W₃, and W₂W₃ is_a W₃, unless an exception is indicated by an operator below.
3. | (bar) indicates that preceding words (up to the latest bar) should be processed independently, such that W₁W₂|W₃ yields the interpretation: W₁W₂ is_a W₂, and W₁W₂W₃ is_a W₂W₃ (e.g., see social media presence)
1. If previous words should be excluded from the concept, then / (slash) can be used to move the previous word marker to the current word position, which excludes previous words from normal processing.
4. > (greater than) indicates that a word should be copied forward one position for each repetition of this operator and independently interpreted with the remaining word sequence, such that W₁>W₂W₃ yields the interpretation: W₁W₂W₃ is_a W₂W₃, W₂W₃ is_a WW₃, and W₁W₃ is_a W₃ (e.g., see real-time geographic location)
   1. If | (bar) precedes (greater than), then the preceeding words will all be copied forward to the to the position indicated, such that W₁W₂|>W₃W₄ yields the interpretation: W₁W₂W₄ is_a W₄ (e.g., see web site usage information)
   2. If a word W₁ should be copied to two or more forward positions, then the | (bar) can be used to introduce an additional copy operation, such that W₁>>|>W₂W₃W₄ yields the interpretation: W₁W₄ is_a W₄ and W₁W₃W₄ is_a W₃W₄ (e.g., anonymous user usage data)
   3. If a word W₁ should be re-located forward, then the / (slash) can be used to exclude the word from normal processing, such that W₁/>W₂|W₃ yields the interpretation: W₁W₃ is_a W₃, but not W₁W₂ is_a W₂ (e.g., see aggregated behavior information).
5. - (hyphen) indicates that the adjacent words should be considered as an inseperable whole when modifying a subsequent word, such that W₁-W₂W₃ yields the interpretation: W₁W₂W₃ is_a W₃, but not W₂W₃ is_a W₃ (e.g., see server provided cookie)
6. _ (underscore) indicates that the matching word(s) are a continuation of the previous word in a multiword phrase that should not be processed normally (e.g., see social media presence, but without the interpretation W₁W₂ is_a W₂)