US 12,292,910 B2
Text classification command line interface tool and related method
Zachary Lloyd Roth Willert, Newton, MA (US); and David James Lustig, Boston, MA (US)
Assigned to Klaviyo, Inc., Boston, MA (US)
Filed by Klaviyo, Inc., Boston, MA (US)
Filed on Feb. 14, 2024, as Appl. No. 18/441,797.
Claims priority of provisional application 63/584,980, filed on Sep. 25, 2023.
Prior Publication US 2025/0103634 A1, Mar. 27, 2025
Int. Cl. G06F 16/335 (2019.01); G06F 9/451 (2018.01); G06F 16/35 (2025.01); G06F 16/383 (2019.01)
CPC G06F 16/335 (2019.01) [G06F 9/453 (2018.02); G06F 16/35 (2019.01); G06F 16/383 (2019.01)] 13 Claims
OG exemplary drawing
 
1. A computer-implemented method for assisting a user with labeling an initial dataset of unlabeled datapoints, the method comprising the steps of:
upon a first computer command, retrieving the initial dataset;
upon a second computer command, filtering the plurality of unlabeled datapoints of the initial dataset to a first set of unlabeled datapoints based on a first keyword associated with a first label;
assigning the first label to at least a portion of the unlabeled datapoints of the first set of unlabeled datapoints after verifying the first label is correct for the unlabeled datapoint being reviewed, thereby generating a first set of verified labeled datapoints;
prompting the user to select a first data-labeling model from a library of pre-existing data-labeling models;
upon a third computer command, selecting the first data-labeling model;
upon a fourth computer command, training, during a first phase, the first data-labeling model based on the first set of verified labeled datapoints; and
upon a fifth computer command, computing a confidence score for each of the datapoints of the initial data set based on the first data-labeling model.