| CPC G06F 16/2246 (2019.01) [G06N 20/00 (2019.01)] | 16 Claims |
|
1. A computer-implemented method for structuring textual data performed by at least one processor, said method comprising:
receiving a description associated with an event;
extracting one or more candidate strings from the description for a respective entity type, wherein the extracting is performed by one or more extractors and each branch score is based on confidence levels of the extractors in the respective branch and an assigned character ratio of the branch, wherein computing a branch score comprises:
computing a harmonic mean of the confidence levels of the extractors in the respective branch;
computing the assigned character ratio of the respective branch; and
combining the harmonic mean and the assigned character ratio to compute the branch score;
outputting a confidence level for each candidate string;
generating a search tree with the one or more candidate strings, the search tree comprising a plurality of branches;
computing a branch score for each of the plurality of branches;
identifying a branch with a highest branch score; and
structuring the description based on the identified branch.
|