US 12,260,303 B2
Machine learning ranking system
Ketakee Kishorkumar Nimavat, Foster City, CA (US); and Rajiv Kumar, Austin, TX (US)
Assigned to Oracle International Corporation, Redwood Shores, CA (US)
Filed by Oracle International Corporation, Redwood Shores, CA (US)
Filed on Feb. 18, 2021, as Appl. No. 17/178,365.
Prior Publication US 2022/0261688 A1, Aug. 18, 2022
Int. Cl. G06F 3/0481 (2022.01); G06F 18/2113 (2023.01); G06F 18/214 (2023.01); G06F 18/22 (2023.01); G06N 20/00 (2019.01)
CPC G06N 20/00 (2019.01) [G06F 3/0481 (2013.01); G06F 18/2113 (2023.01); G06F 18/214 (2023.01); G06F 18/22 (2023.01)] 20 Claims
OG exemplary drawing
 
1. One or more non-transitory computer-readable media storing instructions, which when executed by one or more hardware processors, cause performance of operations comprising:
training a machine learning model to filter and rank entities based on a set of requirements at least by:
obtaining training data sets, each training data set comprising:
attribute values for a set of entity attributes;
a particular set of requirements;
an overall match score representing the match between (a) the attribute values for the set of entity attributes and (b) the particular set of requirements;
a ranking match score representing the match between (a) the attributes values for a first subset of the set of entity attributes and (b) a corresponding first subset of the particular set of requirements;
training the machine learning model based on the training data sets;
applying the machine learning model to a plurality of entities to filter and rank the plurality of entities based on a target set of requirements, the applying generating a ranked subset of entities from the plurality of entities, wherein for each particular entity of the ranked subset of entities:
the machine learning model determines an overall match score based on attribute values of the particular entity for the set of entity attributes and the target set of requirements;
the machine learning model determines a ranking match score representing the match between (a) attributes values of the particular entity for the first subset of the set of entity attributes and (b) a corresponding first subset of the target set of requirements;
the machine learning model determines a rank of the particular entity, among the plurality of entities, based on the ranking match score;
presenting a Graphical User Interface (GUI) comprising a plurality of interface elements respectively representing the ranked subset of entities, wherein the plurality of interface elements are presented in a first order corresponding to the ranking match scores of the respective ranked subset of entities,
wherein the GUI comprises at least one interface feature indicating (a) a proportional contribution to the ranking match score of at least one corresponding attribute and (b) a degree of similarity between the at least one corresponding attribute and at least one corresponding entity attribute value;
receiving user input modifying the first order of the plurality of interface elements to a second order of the plurality of interface elements;
identifying a second subset of the set of entity attributes for computing ranking match scores for the ranked subset of entities that would result in ranking match scores for the ranked subset of entities corresponding to the second order for the respective plurality of interface elements; and
re-training the machine learning model to compute ranking match scores such that the ranking match scores represent the match between (a) attribute values of entities for the second subset of the set of entity attributes and (b) a corresponding second subset of any set of requirements.