System, method, and computer program for user input fields auto-completion using machine learning model selection with dynamic threshold mechanism
As described herein, a system, method, and computer program are provided for using a dynamic threshold mechanism that is utilizing a set of machine learning models to auto-complete user input fields. User access to a form having a plurality of user input fields is detected. One or more of the plurality of user input fields are auto-completed over a sequence of stages, utilizing at least one machine learning model.
1 . A non-transitory computer-readable media storing computer instructions which when executed by one or more processors of a device cause the device to:
detect user access to a form having a plurality of user input fields; and
auto-complete one or more of the plurality of user input fields over a sequence of stages, utilizing at least one machine learning model, wherein a classification threshold used by the at least one machine learning model for each stage in the sequence of stages is adjusted as a function of a current stage in the sequence of stages and F1 scores computed from prior stages in the sequence of stages.
2 . The non-transitory computer-readable media of claim 1 , wherein the device is further caused to:
detect at least one initial user input to at least one user input field of the plurality of user input fields.
3 . The non-transitory computer-readable media of claim 2 , wherein the auto-completing is initiated based on the at least one initial user input to the at least one user input field.
4 . The non-transitory computer-readable media of claim 1 , wherein auto-completing one or more of the plurality of user input fields includes presenting an input suggestion in each of the one or more of the plurality of user input fields.
5 . The non-transitory computer-readable media of claim 4 , wherein the input suggestion is capable of being accepted or rejected by a user.
6 . The non-transitory computer-readable media of claim 5 , wherein acceptance or rejection of the input suggestion affects the auto-completing in at least one subsequent stage in the sequence of stages.
7 . The non-transitory computer-readable media of claim 5 , wherein rejection of the input suggestion presented in a user input field of the plurality of user input fields includes user entry of a new input to the user input field.
8 . The non-transitory computer-readable media of claim 7 , wherein user entry of the new input to the user input field causes the at least one machine learning model to be updated.
9 . The non-transitory computer-readable media of claim 8 , wherein the at least one machine learning model is updated by revising a hierarchy of user input field dependencies.
10 . The non-transitory computer-readable media of claim 1 , wherein auto-completing one or more of the plurality of user input fields over a sequence of stages, utilizing the at least one machine learning model, includes for each stage in the sequence of stages:
processing existing input in the plurality of user input fields, utilizing the at least one machine learning model, to predict additional input for at least one empty user input field of the plurality of user input fields,
auto-completing the at least one empty user input field with the additional input.
11 . The non-transitory computer-readable media of claim 1 , wherein a weight of precision is increased for each subsequent stage in the sequence of stages to increasingly prioritize precision over recall across the sequence of stages.
12 . The non-transitory computer-readable media of claim 1 , wherein the auto-completing is performed utilizing a plurality of machine learning models.
13 . The non-transitory computer-readable media of claim 12 , wherein for each dependent user input field, the plurality of machine learning models are trained using different combinations of independent fields.
14 . The non-transitory computer-readable media of claim 1 , wherein the at least one machine learning model is trained on labeled historical data using supervised learning.
15 . The non-transitory computer-readable media of claim 1 , wherein at least one accuracy metric is calculated for the at least one machine learning model on a validation set.
16 . The non-transitory computer-readable media of claim 15 , wherein the at least one accuracy metric includes one or more of precision, recall, or F1 scores.
17 . The non-transitory computer-readable media of claim 15 , wherein an optimal machine learning model of the at least one machine learning model is selected for each combination of user input fields, based on the at least one accuracy metric, and wherein the optimal machine learning model is utilized for auto-completion involving the associated combination of user input fields.
18 . A method, comprising:
at a computer system:
detecting user access to a form having a plurality of user input fields; and
auto-completing one or more of the plurality of user input fields over a sequence of stages, utilizing at least one machine learning model, wherein a classification threshold used by the at least one machine learning model for each stage in the sequence of stages is adjusted as a function of a current stage in the sequence of stages and F1 scores computed from prior stages in the sequence of stages.
19 . A system, comprising:
a non-transitory memory storing instructions; and
one or more processors in communication with the non-transitory memory that execute the instructions to:
detect user access to a form having a plurality of user input fields; and
auto-complete one or more of the plurality of user input fields over a sequence of stages, utilizing at least one machine learning model, wherein a classification threshold used by the at least one machine learning model for each stage in the sequence of stages is adjusted as a function of a current stage in the sequence of stages and F1 scores computed from prior stages in the sequence of stages.