-
Roles are words, generally noun phrases, that answer questions about a given place. For this tutorial they are:
-
Investor: Who is investing?
-
Money: How much is being invested?
-
Asset: What is being invested in?
-
Key phrases are words, usually but not always verb phrases, used by the model to identify potential events in a piece of text. For this tutorial they include forms of the verb "invest" and the noun "investment."
-
Annotate full tokens, as opposed to subspans of tokens.
-
Minor unintentional misspellings of any named entity should be tagged as usual. This includes: skipped letters, doubled letters, reversed letters, skipped spaces, and inserted letters. Additionally, informal spelling such as repeating the same letter will be tagged as usual.
-
Include punctuation that appears as a part of the named entity or within the named entity.
-
If the named entity is not attached to the adjacent punctuation, and you can highlight and tag that named entity without the adjacent punctuation, do not include the adjacent punctuation that is not a part of the named entity in the tag. This includes trailing periods, quotation marks, parentheses, brackets and all other punctuation that is not part of the named entity.
-
In English, the possessive suffix ‘s is not to be included in the entity, unless it is officially part of the name.
-
Annotations may not overlap or embed in the text. In other words, every annotation must end before another can begin.
-
In English, include "The" in the tag only if it is part of the official name of the named entity. If you are not sure, do a Google search to determine if it is part of the name or not.