US2022156791A1PendingUtilityA1
Method and system for data driven personalization
Est. expirySep 21, 2035(~9.1 yrs left)· nominal 20-yr term from priority
G06Q 30/0255G06Q 30/0277
62
PatentIndex Score
0
Cited by
0
References
0
Claims
Abstract
Systems and methods including one or more processors and one or more non-transitory storage devices storing computing instructions configured to run on the one or more processors and cause the one or more processors to perform tracking, using the one or more processors, usage history of a user; creating training data from the usage history of the user; training a predictive algorithm on the training data; for each respective promotion in a plurality of promotions: assigning a respective weight to the respective promotion of the plurality of promotions by using the predictive algorithm, as trained; after a predetermined time period has elapsed, updating the usage history of the user; re-training the predictive algorithm, as trained, on the usage history of the user, as updated; for each respective promotion in the plurality of promotions: assigning a new respective weight to the respective promotion of the plurality of promotions using the predictive algorithm, as trained and re-trained; and coordinating displaying at least one respective promotion from the plurality of promotions based on the new respective weight, as assigned, for the at least one respective promotion. Other embodiments are disclosed herein.
Claims
exact text as granted — not AI-modifiedWhat is claimed is:
1 . A system comprising:
one or more processors; and one or more non-transitory storage devices storing computing instructions configured to run on the one or more processors and cause the one or more processors to perform:
tracking, using the one or more processors, usage history of a user;
creating training data from the usage history of the user;
training a predictive algorithm on the training data;
for each respective promotion in a plurality of promotions:
assigning a respective weight to the respective promotion of the plurality of promotions by using the predictive algorithm, as trained;
after a predetermined time period has elapsed, updating the usage history of the user;
re-training the predictive algorithm, as trained, on the usage history of the user, as updated;
for each respective promotion in the plurality of promotions:
assigning a new respective weight to the respective promotion of the plurality of promotions using the predictive algorithm, as trained and re-trained; and
coordinating displaying at least one respective promotion from the plurality of promotions based on the new respective weight, as assigned, for the at least one respective promotion.
2 . The system of claim 1 , wherein the predictive algorithm comprises a machine learning algorithm.
3 . The system of claim 1 , wherein the predictive algorithm comprises a maximum entropy model.
4 . The system of claim 1 , wherein the usage history of the user comprises one or more of:
webpage visits of the user; a search history of the user; and a purchase history of the user.
5 . The system of claim 1 , wherein the predictive algorithm is configured to determine a probability of conversion for one or more promotions in the plurality of promotions.
6 . The system of claim 1 , wherein the at least one respective promotion from the plurality of promotions comprises a promotion with a highest probability of conversion.
7 . The system of claim 1 , wherein the training data further comprises location data collected from a GPS module of an electronic device of the user.
8 . The system of claim 1 , wherein using the predictive algorithm comprises using an equation comprising:
argmax
w
∑
t
log
p
(
c
t
|
x
t
,
w
)
=
arg
max
w
∑
t
1
1
+
exp
(
-
w
t
x
)
,
wherein:
w denotes a weight vector;
c denotes a conversion probability;
x denotes a particular feature;
t denotes a time period; and
p denotes an output from a maximum entropy model.
9 . The system of claim 1 , wherein creating the training data comprises:
creating the training data from the usage history of the user and usage history of one or more other users.
10 . The system of claim 1 , wherein the training data serves as a baseline conversion rate for the plurality of promotions.
11 . A method implemented via execution of computing instructions configured to run at one or more processors and configured to be stored at non-transitory computer-readable media, the method comprising:
tracking, using the one or more processors, usage history of a user; creating training data from the usage history of the user; training a predictive algorithm on the training data; for each respective promotion in a plurality of promotions:
assigning a respective weight to the respective promotion of the plurality of promotions by using the predictive algorithm, as trained;
after a predetermined time period has elapsed, updating the usage history of the user; re-training the predictive algorithm, as trained, on the usage history of the user, as updated; for each respective promotion in the plurality of promotions:
assigning a new respective weight to the respective promotion of the plurality of promotions using the predictive algorithm, as trained and re-trained; and
coordinating displaying at least one respective promotion from the plurality of promotions based on the new respective weight, as assigned, for the at least one respective promotion.
12 . The method of claim 11 , wherein the predictive algorithm comprises a machine learning algorithm.
13 . The method of claim 11 , wherein the predictive algorithm comprises a maximum entropy model.
14 . The method of claim 11 , wherein the usage history of the user comprises one or more of:
webpage visits of the user; a search history of the user; and a purchase history of the user.
15 . The method of claim 11 , wherein the predictive algorithm is configured to determine a probability of conversion for one or more promotions in the plurality of promotions.
16 . The method of claim 11 , wherein the at least one respective promotion from the plurality of promotions comprises a promotion with a highest probability of conversion.
17 . The method of claim 11 , wherein the training data further comprises location data collected from a GPS module of an electronic device of the user.
18 . The method of claim 11 , wherein using the predictive algorithm comprises using an equation comprising:
argmax
w
∑
t
log
p
(
c
t
|
x
t
,
w
)
=
arg
max
w
∑
t
1
1
+
exp
(
-
w
t
x
)
,
wherein:
w denotes a weight vector;
c denotes a conversion probability;
x denotes a particular feature;
t denotes a time period; and
p denotes an output from a maximum entropy model.
19 . The system of claim 1 , wherein creating the training data comprises:
creating the training data from the usage history of the user and usage history of one or more other users.
20 . The method of claim 11 , wherein the training data serves as a baseline conversion rate for the plurality of promotions.Join the waitlist — get patent alerts
Track US2022156791A1 — get alerts on status changes and closely related new filings.
We store only your email — no account needed. See our privacy policy.