US 12,014,429 B2
Calibrated risk scoring and sampling
Kate Elizabeth Swift-Spong, San Diego, CA (US); Shivakumara Narayanaswamy, Mountain View, CA (US); Carlos A. Oliveira, Aurora, CO (US); Byungkyu Kang, San Diego, CA (US); Farzaneh Khoshnevisan, San Diego, CA (US); Zhewen Fan, San Diego, CA (US); Runhua Zhao, Milpitas, CA (US); and Wan Yu Zhang, San Jose, CA (US)
Assigned to Intuit Inc., Mountain View, CA (US)
Filed by Intuit Inc., Mountain View, CA (US)
Filed on Jul. 30, 2021, as Appl. No. 17/390,816.
Prior Publication US 2023/0036688 A1, Feb. 2, 2023
Int. Cl. G06Q 20/20 (2012.01); G06F 18/214 (2023.01); G06F 18/243 (2023.01); G06N 20/20 (2019.01); G06Q 10/0635 (2023.01); G06Q 40/12 (2023.01)
CPC G06Q 40/123 (2013.12) [G06F 18/2148 (2023.01); G06F 18/24323 (2023.01); G06N 20/20 (2019.01); G06Q 10/0635 (2013.01); G06Q 20/20 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
extracting features from a record;
generating a risk score, associated with the record, from the features using a machine learning model, wherein the machine learning model is trained to generate risk scores using a plurality of record labels;
mapping the record to a risk bucket of a plurality of risk buckets, using the risk score, wherein the risk bucket comprises a plurality of risk bucket records;
selecting the record, from the plurality of risk bucket records, using a sampling threshold of a plurality of sampling thresholds corresponding to the plurality of risk buckets, wherein the sampling threshold identifies a probability of selecting one of the plurality of risk bucket records from the risk bucket; and
presenting, to a client device, a form using a processor that prepopulates the form with values from the record.