US2023237438A1PendingUtilityA1
Machine learning systems for predictive targeting and engagement
Est. expiryAug 21, 2038(~12.1 yrs left)· nominal 20-yr term from priority
G06Q 10/1053G06N 5/04G06N 20/00G06N 5/046
50
PatentIndex Score
0
Cited by
0
References
0
Claims
Abstract
Machine learning systems for predictive targeting and optimizing engagement are described herein. In various embodiments, the system includes 1) training a first machine learning computer model to generate machine predicted outcomes; (2) determining weights based on the machine predicted outcomes; (3) generating a second machine learning computer model based on the weights; and (4) generating machine learned predictions for candidates.
Claims
exact text as granted — not AI-modifiedWe claim:
1 . A machine learning process for optimized engagement, comprising:
training a primary machine learning model to determine a weight based on a training set; generating a secondary machine learning model based on the weight; generating, via the secondary machine learning model, one or more machine learned predictions; assigning a respective classification to each individual of a plurality of individuals based on the one or more machine learned predictions; processing a set of candidate criteria to identify a subset of individuals from the plurality of individuals that satisfies the set of candidate criteria; determining a ranking of the subset of individuals based on the respective classification assigned to individual ones of the subset of individuals; and generating a notification based on the ranking, wherein the secondary machine learning model is generated by combining one or more intermediary machine learning models according to:
E ( x ijg )= E ( f 1 ( x ijg ), . . . , f n ( x ijg ));
wherein: E(x ijg ) represents the second machine learning model; f is an intermediary machine learning model; x is a vector comprising the normalized data of interest; i is a candidate; j is a company currently employing the candidate i; and g is a role performed by the candidate i at the company j.
2 . The machine learning process of claim 1 , further comprising:
receiving, from one or more databases, data associated with the plurality of individuals, wherein a portion of the received data is received in response to predetermined data triggers configured to monitor changes to a current status of the received data; determining, for each individual of the plurality of individuals, data of interest within the received data, the data of interest comprising candidate information, role information, and company information; normalizing the data of interest into a normalized format to generate normalized data; and creating, from the normalized data of interest, the training set comprising known parameters and known outcomes.
3 . The machine learning process of claim 2 , wherein generating the secondary machine learning model comprises:
creating, from the normalized data of interest, a test set with unknown parameters; and weighting the unknown parameters with the weight.
4 . The machine learning process of claim 3 , further comprising:
calculating the weight from the plurality of machine predicted outcomes and the training set; and weighting the known parameters with the weight via the primary machine learning computer model.
5 . The machine learning process of claim 2 , wherein normalizing the data of interest further comprises performing entity resolution on the data of interest.
6 . The machine learning process of claim 1 , further comprising:
calculating, for each parameter of each identified individual, an impact score; determining, for each parameter and based on each impact score, a most impactful parameter; generating a candidate list based on the ranking of the subset of individuals; and providing, in the candidate list, the most impactful parameter for the individual ones of the subset of individuals.
7 . The machine learning process of claim 6 , wherein determining the most impactful parameter comprises one or more feature importance methods.
8 . The machine learning process of claim 7 , wherein determining the most impactful parameter further comprises:
determining a most positively impactful parameter; and determining a most negatively impactful parameter.
9 . A machine learning process for optimized engagement, comprising:
training a primary one machine learning computer model to determine a weight based a training set; generating a secondary machine learning model based on the weight; generating, via the secondary machine learning computer model, one or more machine learned predictions; generating a talent retention score; and sending a notification based on the talent retention score, wherein the secondary machine learning model is generated by combining one or more intermediary machine learning models according to:
E ( x ijg )= E ( f 1 ( x ijg ), . . . , f n ( x ijg ));
wherein: E(x ijg ) represents the second machine learning model; f is an intermediary machine learning model; x is a vector comprising the normalized data of interest; i is a candidate; j is a company currently employing the candidate i; and g is a role performed by the candidate i at the company j.
10 . The machine learning process of claim 9 , further comprising:
receiving, from one or more databases, data associated with a plurality of employed individuals, wherein a portion of the received data is received in response to predetermined data triggers configured to monitor changes to a current status of the received data; determining, for each individual, data of interest within the received data, the data of interest comprising candidate information, role information, and company information; normalizing the data of interest into a normalized format; and creating, from the normalized data, the training set comprising known parameters and known outcomes.
11 . The machine learning process of claim 10 , wherein the plurality of individuals are employed in a particular department.
12 . The machine learning process of claim 10 , wherein the plurality of individuals are employed in a particular location.
13 . The machine learning process of claim 10 , wherein the plurality of individuals present a particular experience level.
14 . The machine learning process of claim 10 , wherein generating the secondary machine learning model comprises:
creating, from the normalized data of interest, a test set with unknown parameters; and weighting the unknown parameters with the weight.
15 . The machine learning process of claim 9 , further comprising:
comparing the talent retention score to one or more predefined thresholds; and based on the comparison, assigning a classification to the talent retention score, wherein the classification is selected from a group comprising of low, average, and high.
16 . A machine learning system for optimized engagement, comprising:
a database; and a processor in communication with the database, the processor being configured to: generate a machine learning model based on a weight based on a machine predicted outcome and a training set; generate, via the machine learning model, one or more machine learned predictions; and assign a respective classification to each individual of a plurality of individuals based on the one or more machine learned predictions, wherein the processor, to generate the machine learning model, is further configured to generate the machine learning model by combining one or more intermediary machine learning models according to:
E ( x ijg )= E ( f 1 ( x ijg ), . . . , f n ( x ijg ));
wherein: E(x ijg ) represents the second machine learning model; f is an intermediary machine learning model; x is a vector comprising the normalized data of interest; i is a candidate; j is a company currently employing the candidate i; and g is a role performed by the candidate i at the company j.
17 . The machine learning system of claim 16 , wherein the processor is further configured to:
receive one or more predetermined triggers configured to monitor changes to a current status of data associated with the plurality of individuals, wherein the one or more predetermined triggers cause the processor to retrieve, from the database, the data associated with the plurality of individuals; determine a respective data of interest within the retrieved data for each of the plurality of individuals, the respective data of interest comprising candidate information, role information, and company information; normalize the respective data of interest for each of the plurality of individuals into a normalized format; and create the training set from the normalized data, the training set comprising known parameters and known outcomes.
18 . The machine learning system of claim 17 , wherein:
the candidate information comprises current tenure, average tenure in previous roles, number of previous roles with current company, number of previous roles at other companies, skills, education level, relative pay, previous industries, previous company size, previous company age, geography, and commute time; the role information comprises title, level, functions, similar open positions, and open growth opportunities; and the company information comprises company type, company size, age, brand measurements, news and events, and trends in news and events.
19 . The machine learning system of claim 18 , wherein the company type is selected from a group comprising: public, private, government, and school.
20 . The machine learning system of claim 17 , wherein:
the processor is further configured to train, with the training set, a primary machine learning computer model to generate the machine predicted outcome, wherein the processor is configured to train the primary machine learning computer model by: calculating, from the machine predicted outcome and the training set, the weight; and weighting the known parameters with the weight and via the primary machine learning computer model.
21 . The machine learning system of claim 16 , wherein generating the machine learning model comprises:
creating, from normalized data of interest, a test set with unknown parameters; and weighting the unknown parameters with the weight.
22 . The machine learning system of claim 16 , wherein the processor is further configured to:
compare the one or more machine learned predictions and the classifications to third-party data from the database; determine particular language comprising subject lines and keywords from the third-party data that is likely to elicit a response from each of the plurality of individuals, based on the one or more machine learned predictions and the classifications; and generate one or more strings of text via natural language processing, wherein the one or more strings of text comprise language substantially similar to the particular language.
23 . The machine learning system of claim 22 , wherein the processor, to generate the one or more strings of text, is further configured to select at least a portion of the one or more strings of text using a conditional logic process.
24 . The machine learning system of claim 16 , wherein the processor, to assign the classification, is configured to evaluate and assign each of the one or more machine learned predictions according to:
c
(
x
ijg
)
=
{
candidate
is
least
likely
to
ENGAGE
if
h
(
x
i
j
g
)
<
h
0
candidate
may
ENGAGE
if
h
0
<
h
(
x
i
j
g
)
<
h
1
candidate
is
more
likely
to
ENGAGE
h
1
<
h
(
x
i
j
g
)
<
h
2
candidate
is
most
likely
to
ENGAGE
h
(
x
i
j
g
)
>
h
2
wherein:
h(x ijg ) is a machine learned prediction from the one or more machine learned predictions;
h 0 is a predefined will-not-engage threshold;
h 1 is a predefined may-engage threshold;
h 2 is a predefined more-likely-to-engage threshold; and
c(x ijg ) is the classification to which each one the one or more machine learned predictions is assigned.
25 . The machine learning system of claim 16 , wherein the processor is further configured to:
retrieve, from the database, a plurality of historical classifications associated with the plurality of individuals; determine, for each individual, if a historical classification matches the assigned classification; upon determining, for a particular individual, that the assigned classification does not match the historical classification, determine if the particular individual is included on a recruitment watch list; and upon determining, that the particular individual is included on the recruitment watch list, automatically generate and transmit, to a profile associated with the recruitment watch list, an alert describing that a classification for the particular individual has changed.
26 . The machine learning system of claim 25 , wherein the assigned classification of the particular individual is determined to exceed the historical classification.
27 . The machine learning system of claim 16 , wherein the processor is further configured to generate, for each individual, a data visualization comprising the classification and the one or more machine learned predictions.
28 . The machine learning system of claim 27 , wherein the data visualization is a radar chart.Join the waitlist — get patent alerts
Track US2023237438A1 — get alerts on status changes and closely related new filings.
We store only your email — no account needed. See our privacy policy.