# US 2023/0409773 A1 · Reach and frequency prediction for digital component transmissions

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2023/0409773 A1
- Document ID: US20230409773A1
- Status: Published application
- Kind: A1
- Application number: 17/845,827
- Filed: 2022-06-21
- Published: 2023-12-21
- Earliest claimed: 2022-06-21
- Assignees: Google Inc.
- Applicants: Google LLC
- Inventors: Jiayu Peng; Matthew Tran Clegg
- Web page: https://patentgrep.com/patent/US20230409773A1

## CPC

- G06F 30/20: Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30)
- G06F 2111/08: Probabilistic or stochastic CAD
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

In one aspect, there is provided a method performed by one or more computers that includes: obtaining an observed frequency histogram corresponding to an observed transmission commitment, where a transmission commitment specifies a number of transmissions of a digital component; generating a frequency model based on the observed frequency histogram, where the frequency model is a parametric model parameterized by a set of model parameters; receiving a request to predict a frequency histogram corresponding to a target transmission commitment; processing data defining the target transmission commitment using the frequency model to generate a predicted frequency histogram corresponding to the target transmission commitment; and generating one or more predictions characterizing the target transmission commitment using the predicted frequency histogram.

## Claims

### Claim 1

A method performed by one or more computers, the method comprising:
obtaining an observed frequency histogram corresponding to an observed transmission commitment, wherein:
a transmission commitment specifies a number of transmissions of a digital component, and
a frequency histogram corresponding to the transmission commitment defines, for each of a plurality of frequency values, a respective number of users who each receive a number of transmissions of the digital component given by the frequency value when the digital component is transmitted a number of times specified by the transmission commitment;
generating a frequency model based on the observed frequency histogram, wherein:
the frequency model is a parametric model parameterized by a set of model parameters,
the frequency model is configured to process an input defining an input transmission commitment to generate an output that defines a predicted frequency histogram corresponding to the input transmission commitment, and
generating the frequency model comprises training the set of model parameters of the frequency model, using a numerical optimization technique, to optimize an objective function that depends on the observed frequency histogram;
receiving a request to predict a frequency histogram corresponding to a target transmission commitment; and
in response to receiving the request:
processing data defining the target transmission commitment using the frequency model to generate a predicted frequency histogram corresponding to the target transmission commitment; and
generating one or more predictions characterizing the target transmission commitment using the predicted frequency histogram.

### Claim 2 (depends on 1)

The method of claim 1, wherein the frequency model comprises a parametric probability distribution over the plurality of frequency values.

### Claim 3 (depends on 2)

The method of claim 2, wherein the parametric probability distribution over the plurality of frequency values is a shifted Gamma Poisson distribution.

### Claim 4 (depends on 2)

The method of claim 2, wherein the parametric probability distribution over the plurality of frequency values is a k-inflated Gamma Poisson distribution.

### Claim 5 (depends on 2)

The method of claim 2, wherein for an input transmission commitment specifying Ninp transmissions of the digital component, the frequency model generates a predicted frequency histogram that, for each of one or more frequencies i, defines a predicted number of users Ri corresponding to frequency i as:
R
i
=
N
pop
⁢
∑
i
≤
j
≤
m
(
j
i
)
·
(
N
inp
N
inv
)
i
·
(
1
-
N
inp
N
inv
)
j
-
i
·
f
⁡
(
j
⁢
❘
"\[LeftBracketingBar]"
θ
)
wherein Npop defines a total number of users, m defines a total number of frequency values, j indexes frequency values between i and m, Ninv, is a model parameter of the frequency model defining an inventory-level number of transmissions of the digital component, f(j|θ) is the parametric probability distribution over the plurality of frequency values evaluated at frequency j, and θ are model parameters of the frequency model that parameterize the probability distribution f(⋅|θ).

### Claim 6 (depends on 1)

The method of claim 1, wherein the objective function measures an error between: (i) a predicted frequency histogram generated by processing data defining the observed transmission commitment using the frequency model, and (ii) the observed frequency histogram.

### Claim 7 (depends on 6)

The method of claim 6, wherein the objective function measures the error using a χ2 statistic.

### Claim 8 (depends on 6)

The method of claim 6, wherein the objective function comprises a penalty term that encourages an inventory-level number of transmissions of the digital component to exceed a multiple of the observed number of transmissions of the digital component that is specified by the observed transmission commitment.

### Claim 9 (depends on 8)

The method of claim 8, wherein the multiple is greater than 2.

### Claim 10 (depends on 9)

The method of claim 9, wherein the objective function is given by:
ℒ
=
∑
i
=
1
f
(
h
i
-
h
ˆ
i
)
2
h
ˆ
i
+
(
sN
obs
-
N
inv
)
[
sN
obs
>
N
inv
]
wherein i∈{1, . . . , f} indexes the plurality of frequency values, hi is a number of users corresponding to frequency i under the observed frequency histogram, ĥi is a predicted number of users corresponding to frequency i under the predicted frequency histogram generated by processing data defining the observed transmission commitment using the frequency model, s>1 is a scalar multiple, Nobs is the observed number of transmissions of the digital component, Ninv is an inventory-level number of transmissions of the digital component, and [⋅] is an indicator function.

### Claim 11 (depends on 1)

The method of claim 1, wherein the numerical optimization technique is a Nelder-Mead numerical optimization technique.

### Claim 12 (depends on 1)

The method of claim 1, wherein generating one or more predictions characterizing the target transmission commitment using the predicted frequency histogram comprises:
generating a prediction for a number of users that will receive a specified number of transmissions of the digital component under the target transmission commitment.

### Claim 13 (depends on 1)

The method of claim 1, wherein generating one or more predictions characterizing the target transmission commitment using the predicted frequency histogram comprises:
generating a prediction for a total number of users that will receive at least one transmission of the digital component under the target transmission commitment.

### Claim 14

A system comprising:
one or more computers; and
one or more storage devices communicatively coupled to the one or more computers, wherein the one or more storage devices store instructions that, when executed by the one or more computers, cause the one or more computers to perform operations comprising:
obtaining an observed frequency histogram corresponding to an observed transmission commitment, wherein:
a transmission commitment specifies a number of transmissions of a digital component, and
a frequency histogram corresponding to the transmission commitment defines, for each of a plurality of frequency values, a respective number of users who each receive a number of transmissions of the digital component given by the frequency value when the digital component is transmitted a number of times specified by the transmission commitment;
generating a frequency model based on the observed frequency histogram, wherein:
the frequency model is a parametric model parameterized by a set of model parameters,
the frequency model is configured to process an input defining an input transmission commitment to generate an output that defines a predicted frequency histogram corresponding to the input transmission commitment, and
generating the frequency model comprises training the set of model parameters of the frequency model, using a numerical optimization technique, to optimize an objective function that depends on the observed frequency histogram;
receiving a request to predict a frequency histogram corresponding to a target transmission commitment specifying a target number of transmissions of the digital component; and
in response to receiving the request:
processing data defining the target transmission commitment using the frequency model to generate a predicted frequency histogram corresponding to the target transmission commitment; and
generating one or more predictions characterizing the target transmission commitment using the predicted frequency histogram.

### Claim 15

One or more non-transitory computer storage media storing instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:
obtaining an observed frequency histogram corresponding to an observed transmission commitment, wherein:
a transmission commitment specifies a number of transmissions of a digital component, and
a frequency histogram corresponding to a transmission commitment defines, for each of a plurality of frequency values, a respective number of users who each receive a number of transmissions of the digital component given by the frequency value when the digital component is transmitted a number of times specified by the transmission commitment;
generating a frequency model based on the observed frequency histogram, wherein:
the frequency model is a parametric model parameterized by a set of model parameters,
the frequency model is configured to process an input defining an input transmission commitment to generate an output that defines a predicted frequency histogram corresponding to the input transmission commitment, and
generating the frequency model comprises training the set of model parameters of the frequency model, using a numerical optimization technique, to optimize an objective function that depends on the observed frequency histogram;
receiving a request to predict a frequency histogram corresponding to a target transmission commitment specifying a target number of transmissions of the digital component; and
in response to receiving the request:
processing data defining the target transmission commitment using the frequency model to generate a predicted frequency histogram corresponding to the target transmission commitment; and
generating one or more predictions characterizing the target transmission commitment using the predicted frequency histogram.

### Claim 16 (depends on 15)

The non-transitory computer storage media of claim 15, wherein the frequency model comprises a parametric probability distribution over the plurality of frequency values.

### Claim 17 (depends on 16)

The non-transitory computer storage media of claim 16, wherein the parametric probability distribution over the plurality of frequency values is a shifted Gamma Poisson distribution.

### Claim 18 (depends on 16)

The non-transitory computer storage media of claim 16, wherein the parametric probability distribution over the plurality of frequency values is a k-inflated Gamma Poisson distribution.

### Claim 19 (depends on 16)

The non-transitory computer storage media of claim 16, for an input transmission commitment specifying Ninp transmissions of the digital component, the frequency model generates a predicted frequency histogram that, for each of one or more frequencies i, defines a predicted number of users Ri corresponding to frequency i as:
R
i
=
N
pop
⁢
∑
i
≤
j
≤
m
(
j
i
)
·
(
N
inp
N
inv
)
i
·
(
1
-
N
inp
N
inv
)
j
-
i
·
f
⁡
(
j
⁢
❘
"\[LeftBracketingBar]"
θ
)
wherein Npop defines a total number of users, m defines a total number of frequency values, j indexes frequency values between i and m, Ninv is a model parameter of the frequency model defining an inventory-level number of transmissions of the digital component, f(j|θ) is the parametric probability distribution over the plurality of frequency values evaluated at frequency j, and θ are model parameters of the frequency model that parameterize the probability distribution f(⋅|θ).

### Claim 20 (depends on 15)

The non-transitory computer storage media of claim 15, wherein the objective function measures an error between: (i) a predicted frequency histogram generates by processing data defining the observed transmission commitment using the frequency model, and (ii) the observed frequency histogram.
