US 12,223,397 B2
Action determination using recommendations based on prediction of sensor-based systems
Amit Vaid, Bengaluru (IN); and Vijayalakshmi Krishnamurthy, Sunnyvale, CA (US)
Assigned to Oracle International Corporation, Redwood Shores, CA (US)
Filed by Oracle International Corporation, Redwood Shores, CA (US)
Filed on Aug. 31, 2020, as Appl. No. 17/007,601.
Prior Publication US 2022/0067572 A1, Mar. 3, 2022
Int. Cl. G06N 5/022 (2023.01); G06F 16/28 (2019.01); G06F 18/214 (2023.01); G06F 40/143 (2020.01); G06F 40/35 (2020.01); G06F 40/58 (2020.01); G06N 3/042 (2023.01); G06N 20/00 (2019.01)
CPC G06N 20/00 (2019.01) [G06F 16/284 (2019.01); G06F 18/214 (2023.01)] 21 Claims
OG exemplary drawing
 
1. One or more non-transitory computer readable media comprising instructions which, when executed by one or more hardware processors, cause performance of operations comprising:
generating a first machine learning model to predict performance values for a device, at least by:
obtaining first training datasets comprising a plurality of data points, each training dataset comprising:
a data point comprising a set of parameter values corresponding to a set of operation parameters of the device; and
a performance value corresponding to a performance of the device;
training the first machine learning model based on the first training datasets to predict the performance values for the device;
applying a first data point, comprising a first set of values for a set of parameters for a target device, to the first machine learning model to predict a first performance value of the target device associated with the first data point;
based at least on the first performance value meeting a selection criteria, mapping the first data point in a multi-dimensional space;
generating, in the multi-dimensional space, a multi-dimensional boundary in relation to the first data point based on at least a multi-dimensional distance of a first set of data points, defining the multi-dimensional boundary, from the first data point;
selecting a second set of data points within the multi-dimensional boundary;
generating a second training dataset from the second set of data points;
generating a second machine learning model based on training the second machine learning model with the second training dataset;
selecting a third set of data points within the multi-dimensional boundary as candidate data points;
for each particular candidate data point of the third set of data points:
applying the particular candidate data point to the second machine learning model to predict a particular performance value of the target device corresponding to the particular candidate data point;
identifying a subset of the candidate data points (a) with corresponding performance values that meet performance criteria and (b) that are within a threshold range from the first data point in the multi-dimensional space; and
recommending, for the operation parameters of the target device, at least one candidate set of values corresponding to at least one candidate data point (a) with one or more corresponding performance values that meet the performance criteria and (b) that is within the threshold range from the first data point in the multi-dimensional space.