US 12,353,492 B1
Machine learning system and method for targeted growth projection of market inquiry results
Jennifer Brussow, Syracuse, NY (US); Patrick Danial, Syracuse, NY (US); and Christopher Pasquarella, Syracuse, NY (US)
Assigned to Terrakeet LLC, Syracuse, NY (US)
Filed by Terakeet LLC, Syracuse, NY (US)
Filed on Dec. 24, 2024, as Appl. No. 19/001,083.
Application 19/001,083 is a continuation in part of application No. 18/654,419, filed on May 3, 2024, granted, now 12,182,217.
Int. Cl. G06F 16/9538 (2019.01); G06F 16/955 (2019.01)
CPC G06F 16/9538 (2019.01) [G06F 16/9566 (2019.01)] 14 Claims
OG exemplary drawing
 
1. A method of predicting uniform resource locator (URL) positioning, comprising:
generating, by a computer system, for one or more identified keywords obtained from one or more external resources, a plurality of corresponding URLs for search engine result pages (SERPs);
determining, by the computer system, one or more classifications of each URL of the plurality of corresponding URLs, wherein the one or more classifications comprise branding classification;
filtering, by the computer system, the plurality of corresponding URLs by the one or more classifications to exclude branding classification for a respective URL;
generating, by the computer system, a filtered set of the filtered plurality of corresponding URLs based on at least the excluded branding classification of the one or more classifications;
selecting, from the filtered set of the plurality of corresponding URLs, a subset of candidate URLs for optimization, wherein the selecting candidate URLs for optimization is based on URLs of the filtered set of the plurality of corresponding URLs, and wherein a candidate URL of the selected candidate URLs comprises a monthly search volume (MSV);
determining, by the computer system, a plurality of URL characteristics corresponding to the filtered set of the plurality of corresponding URLs;
in response to a request from a client device, generating, by the computer system, a prediction of SERP positioning of the candidate URL of the selected candidate URLs for the one or more identified keywords by converting the plurality of URL characteristics into input for a machine learning model executed on the computer system;
applying, by the computer system, the input to the machine learning model to evaluate the prediction of SERP positioning for a predetermined time horizon corresponding respective estimates of website traffic, and in response obtaining, for the predetermined time horizon measured from a current month, predicted SERP positions of the candidate URL of the selected candidate URLs for the one or more identified keywords; and
providing, via an application interface of the computer system, an entity access to the predicted SERP positions of the candidate URL of the selected candidate URLs and display of the predicted SERP positions and the respective estimates of website traffic.