Systems and methods for classifying a webpage or a webpage element

1 . A computer-implemented method for classifying a webpage or one or more webpage elements, comprising:

detecting, by an electronic application operating on a user device, that a user of the user device has navigated to the webpage using a web browser also operating on the user device;

extracting, by the electronic application, first input data from the web browser, wherein the first input data includes at least:

a document object model (DOM) of the webpage,

text associated with the webpage or the one or more webpage elements, and

a position of the one or more webpage elements as rendered on a display associated with the user device;

classifying, by the electronic application, the webpage or the one or more webpage elements as either (i) transaction-related or (ii) not transaction-related, by inputting the first input data into one or more machine learning models of the electronic application that operate locally on the user device, wherein:

the one or more machine learning models have been trained based on one or more pre-classified webpages or webpage elements so as to learn associations between the one or more pre-classified webpages or webpage elements and one or more classifications, such that the one or more machine learning models are configured to perform the classifying based on the learned associations; and

classifying the webpage is based on one or more fields included in the DOM, one or more attributes associated with the one or more fields, or a pixel position of the one or more webpage elements as rendered on the display;

in response to classifying the webpage or the one or more webpage elements as transaction-related, causing the user device to display a user interface associated with the electronic application, the user interface including information identifying the classifying;

generating, by the electronic application and based on the classification of the webpage or the one or more webpage elements, one or more metrics related to the classifying;

transmitting, by the electronic application, the one or more metrics to a server that is separate from the user device; and

receiving, via the electronic application and from the server, one or more updated machine learning models that have been retrained based on the one or more metrics.

2 . The computer-implemented method of claim 1 ,

wherein the user interface further includes one or more input controls operable to cause the electronic application to populate transaction card-related data into at least one of the one or more webpage elements, and

wherein the method further comprises:

based on selection of the one or more input controls, populating the transaction card-related data into the at least one of the one or more webpage elements.

3 . The computer-implemented method of claim 1 , wherein the one or more machine learning models comprise:

a first machine learning model trained to classify the webpage based on the text, and

a second machine learning model trained to classify the one or more elements based on the DOM, the text, and the position.

4 . The computer-implemented method of claim 1 , further comprising:

performing a featurization of data associated with the webpage prior to classifying the webpage or the one or more webpage elements.

5 . The computer-implemented method of claim 1 , wherein the electronic application comprises an extension or a plugin installed on the web browser.

6 . The computer-implemented method of claim 1 , wherein the classifying of the webpage comprises:

classifying the webpage based on:

the text in a title or a header of the webpage, or

the text in labels associated with input elements or selection elements of the webpage.

7 . The computer-implemented method of claim 1 , wherein the classifying of the one or more webpage elements comprises:

classifying the one or more webpage elements based on:

one or more fields included in the DOM,

one or more attributes for the one or more fields in the DOM, or

a relationship between the text and the one or more webpage elements based on a pixel position of the text and a pixel position of the one or more webpage elements.

8 . The computer-implemented method of claim 1 , wherein the classifying further comprises:

determining a first score for the webpage and a second score for the one or more webpage elements, wherein the first score and the second score indicate a fit of the webpage or the one or more webpage elements, respectively, to the classification; and

classifying the webpage and the one or more webpage elements based on the first score and the second score.

9 . The computer-implemented method of claim 1 , wherein the classifying further comprises:

determining that the webpage is transaction-related based on the text in a title or a header of the webpage, text associated with input webpage elements or selection webpage elements of the one or more webpage elements, and particular terms included in the text.

10 . A computer-implemented method for training a machine-learning model for classifying one or more webpages or webpage elements, comprising:

receiving data associated with one or more webpages or one or more webpage elements,

wherein the one or more webpages or the one or more webpage elements are labelled with a classification of either (i) transaction-related or (ii) not transaction-related,

wherein the data comprises document object models (DOMs) of the one or more webpages, one or more fields included in the DOMs, one or more attributes associated with the one or more fields, text associated with the one or more webpages or the one or more webpage elements, and pixel positions of the one or more webpage elements as rendered on a display associated with a user device;

processing the data to extract one or more features from the data;

training one or more machine learning models to classify one or more other webpages or one or more other webpage elements based on the extracted features;

providing a software module for installation on one or more user devices, wherein the software module comprises the one or more trained machine learning models and an electronic application configured to use the one or more machine learning models;

receiving, from the one or more user devices, one or more metrics generated by the electronic application of each user device of the one or more user devices, the one or more metrics related to classifying of webpages as either (i) transaction related or (ii) not transaction related;

retraining the one or more machine learning models based on the one or more metrics from the one or more user devices; and

providing an updated software module for installation on the one or more user devices, wherein the updated software module comprises the one or more retrained machine learning models.

11 . The computer-implemented method of claim 10 ,

wherein the training of the one or more machine learning models further comprises:

training the one or more machine learning models using a first programming language; and

wherein the method further comprises:

transforming the one or more machine learning models from the first programming language to a second programming language.

12 . The computer-implemented method of claim 10 , wherein the processing of the data to extract the one or more features further comprises:

processing the data using at least a term frequency-inverse document frequency analysis and a logistic regression analysis.

13 . The computer-implemented method of claim 10 , wherein the training of the one or more machine learning models further comprises:

determining a first score that indicates an accuracy of the classification of the one or more webpages; and

determining a second score that indicates an accuracy of the classification of the one or more webpage elements.

14 . The computer-implemented method of claim 13 , further comprising:

performing a model weights analysis based on the first score and the second score.

15 . The computer-implemented method of claim 10 , wherein the providing of the software module further comprises:

providing the software module to a server device for installation by the user device, or

installing the software module in a web browser on the user device as an extension or a plugin.

16 . The computer-implemented method of claim 10 , wherein the one or more machine learning models each comprise one or more neural networks with a single node.

17 . A system for classifying a webpage or one or more webpage elements, comprising:

a server device configured to receive one or more metrics and retrain one or more machine learning models based on the received one or more metrics; and

a user device that includes:

a memory storing:

instructions; and

an electronic application that includes one or more machine learning models trained, based on webpages and associated webpage elements, to learn associations between the webpages and associated webpage elements and output classifications based on the learned associations in response to input webpage-related data, wherein the one or more machine learning models operate locally on the user device; and

a processor operatively connected to the memory and configured to execute the instructions to perform operations including:

receiving, by the electronic application, an indication to start monitoring web browsing activity of a user of the user device;

detecting, by the electronic application, that the user has navigated to the webpage using a web browser installed on the user device;

extracting, by the electronic application, first input data from the web browser, wherein the first input data includes at least:

a document object model (DOM) of the webpage,

one or more fields included in the DOM,

one or more attributes associated with the one or more fields,

text associated with the webpage or the one or more webpage elements, and

a pixel position of the one or more webpage elements as rendered on a display associated with the user device;

classifying, by the electronic application, the webpage or the one or more webpage elements as either (i) transaction-related or (ii) not transaction related, by inputting the first input data into the one or more machine learning models,

in response to classifying the webpage or the one or more webpage elements as transaction-related, causing the user device to display a user interface associated with the electronic application, the user interface including:

information identifying the classifying, and

one or more input controls associated with causing the electronic application to populate transaction-related data into at least one of the one or more webpage elements;

based on selection of the one or more input controls, populating, by the user device, the transaction-related data into the at least one of the one or more elements;

generating, by the electronic application and based on the selection of the one or more input controls, one or more metrics related to the classifying;

transmitting, by the electronic application, the one or more metrics to the server; and

receiving, via the electronic application and from the server, one or more updated machine learning models that have been retrained based on the one or more metrics.

18 . The computer-implemented method of claim 1 , wherein the one or more updated machine learning models have been retrained based on one or more metrics from a plurality of different user devices.

19 . The computer-implemented method of claim 10 , wherein the retraining of the one or more machine learning models is based on one or more metrics from a plurality of user devices.