| 主权项 |
1. A computer-implemented method of processing natural language for machine learning analytical systems, the method comprising:
receiving, by a processor, tracked observed parameters of agent behavior input by a tracking system including at least one of a satellite and a radar system; receiving, by a natural language (NL) module, an input including text representing one or more observed parameters of an environment, the inputted text being in a natural language format; using the NL module to parse the input in the natural language format and extract at least one of a spatial concept and a temporal concept from the inputted text; defining, by the NL module, a mathematical function f(s,t) of space and time representing a domain of the one or more observed parameters based upon the at least one of the spatial concept and the temporal concept, wherein the mathematical function f(s,t) returns a value of approximately 1.0 when s describes spatial proximity to the spatial concept and t describes a temporal proximity to the temporal concept, and declines toward 0.0 as the space and time supplied deviate from the spatial and temporal concepts; storing the defined mathematical function f(s,t) as a domain descriptor in a database communicatively coupled to the NL module; calculating, by the processor, a likelihood that the tracked observed parameters from the tracking system correlate to the at least one of the spatial concept and the temporal concept based on the mathematical function f(s,t); and displaying, by the processor, an alert on a display based on the calculated likelihood. |
