US 11,973,841 B2
System and method for user model based on app behavior
Ayman Farahat, Santa Clara, CA (US); and Tarun Bhatia, Burbank, CA (US)
Assigned to Yahoo Ad Tech LLC, New York, NY (US)
Filed by Yahoo Holdings, Inc., Sunnyvale, CA (US)
Filed on Dec. 29, 2015, as Appl. No. 14/982,997.
Prior Publication US 2017/0185901 A1, Jun. 29, 2017
Int. Cl. G06N 5/04 (2023.01); G06F 16/00 (2019.01); G06F 16/9535 (2019.01); H04L 67/50 (2022.01); H04L 67/53 (2022.01)
CPC H04L 67/535 (2022.05) [G06F 16/00 (2019.01); G06F 16/9535 (2019.01); G06N 5/04 (2013.01); H04L 67/53 (2022.05)] 17 Claims
OG exemplary drawing
 
1. A system for building a user model, comprising:
a processor and a non-transitory storage medium accessible to the processor, the processor configured to:
obtain user data from a database, wherein the user data comprise user behavior for a plurality of apps installed on one or more user terminals;
select at least one rating parameter using the user data, wherein the at least one rating parameter indicates an explicit rating, assigned by a user in an app store, of at least one app in the app store;
build the user model based on a user-to-rating matrix comprising the at least one rating parameter, wherein the user-to-rating matrix comprises a first dimension indicating a plurality of users of the at least one app in the app store and a second dimension indicating the plurality of apps, wherein the building the user model comprises implementing a matrix factorization algorithm that uses Alternating Least Squares with Weighted-Lambda-Regularization (ALS-WR) to factor the user-to-rating matrix into a user-to-feature matrix and a rating-to-feature matrix;
estimate app usage using the user model built based on the user-to-rating matrix indicating the plurality of users of the at least one app in the app store; and
provide a recommendation of two or more candidate users, of the plurality of users of the at least one app in the app store that are indicated in the user-to-rating matrix used to build the user model used to estimate the app usage, to one or more app developers based on the app usage, wherein the recommendation comprises a first selectable option corresponding to a first candidate user of the two or more candidate users and a second selectable option corresponding to a second candidate user of the two or more candidate users.