CPC G06N 3/08 (2013.01) [G06F 16/3344 (2019.01); G06F 40/253 (2020.01); G06F 40/30 (2020.01); G06N 3/02 (2013.01); G06N 3/045 (2023.01); G10L 15/16 (2013.01); G06Q 10/107 (2013.01)] | 20 Claims |
1. A method for analyzing grammar in a natural language message, comprising:
receiving a natural language message having a type a model of hierarchical complexity level of at least 6;
detecting an ordered set of words within the natural language message;
linking the set of words found within the natural language message to a corresponding set of expected words, the set of expected words having semantic attributes, the corresponding set of expected words being dependent on natural language training data;
detecting a set of grammatical structures represented in the natural language message, based on:
the type of the received natural language message,
the ordered set of words, and
the semantic attributes of the corresponding set of expected words; and
producing at least one output selectively dependent on the set of grammatical structures represented in the natural language message.
|