Method and an apparatus for displaying information flow, a storage, an electronic device

Patent metadata

Field	Value
Publication number	US-11334707-B2
Application number	US-201816979690-A
Country	US
Kind code	B2
Filing date	Nov 14, 2018
Priority date	Nov 14, 2018
Publication date	May 17, 2022
Grant date	May 17, 2022

Abstract

A method and an apparatus for displaying information flow, a storage medium and an electronic device are provided. The method includes: dividing a predetermined display time period for the first-type display information into multiple sub-time periods; determining, for the sub-time periods, a segmented display target of each piece of first-type display information in the sub-period based on a display target of the first-type display information in the predetermined display time period; and determining, for each display position, target display information to be allocated to the display position from the information flow, based on an estimated click-through rate of the display position, a proposed value of the second-type display information, historical display data of the first-type display information, and a segmented display target of the first-type display information in a current sub-time period.

First claim

The invention claimed is:

1. A method for displaying information flow, wherein the information flow comprises a plurality of pieces of first-type display information and a plurality of pieces of second-type display information, wherein the first-type display information is associated with a predetermined display time period of displaying the plurality of pieces of the first-type display information, and each of the plurality of pieces of the first-type display information corresponds to a predetermined number of displaying each piece in the predetermined display time period, wherein each piece of the second-type display information is associated with a value that is obtained by displaying the second-type display information, wherein displaying the first-type display information instead of the second-type display information causes a value loss, and wherein the method comprises:
dividing the predetermined display time period for the plurality of pieces of first-type display information into a plurality of sub-time periods;
determining, for each of the plurality of sub-time periods, a segmented display target of each of the plurality of pieces of the first-type display information in each sub-time period based on the predetermined number for displaying each of the plurality of pieces of the first-type display information in the predetermined display time period, wherein the segmented display target comprises a target number of displaying each of the plurality of pieces of the first-type display information in each of the plurality of sub-time periods; and
determining, for each display position, target display information to be allocated to the display position from the information flow, based on an estimated click-through rate of the display position, the value of each of the plurality of pieces of the second-type display information, historical display data of each of the plurality of pieces of the first-type display information, and the segmented display target of each of the plurality of pieces of the first-type display information in a current sub-time period.

2. The method according to claim 1 , wherein the determining, for each of the plurality of sub-time periods, a segmented display target of each of the plurality of pieces of the first-type display information in the sub-time period based on the predetermined number for displaying each of the plurality of pieces of the first-type display information in the predetermined display time period further comprises:
obtaining, for each of the plurality of sub-time periods, a cumulative display number and a cumulative click-through number of each of the plurality of pieces of the first-type display information at a starting time of each sub-time period;
determining a total return of each piece of the first-type display information in each sub-time period using a segment target determination model based on state information associated with each sub-time period and each piece of the first-type display information, wherein the state information comprises the starting time of each sub-time period, the cumulative display number and the cumulative click-through number of each piece of the first-type display information at the starting time, and wherein the segment target determination model is pretrained by performing reinforcement learning using a large amount of historical first-type display information;
determining a target conversion rate of each of the plurality of pieces of the first-type display information in each sub-time period based on the total return; and
determining, for each of the plurality of pieces of the first-type display information, the target number of displaying each piece of the first-type display information in each sub-time period based on the predetermined number of displaying each piece of the first-type display information in the predetermined display time period and the target conversion rate of each piece of the first-type display information in each sub-time period.

3. The method according to claim 1 , wherein the determining, for each display position, target display information to be allocated to the display position from the information flow, based on an estimated click-through rate of the display position, the value of each of the plurality of pieces of the second-type display information, historical display data of each of the plurality of pieces of the first-type display information, and the segmented display target of each of the plurality of pieces of the first-type display information in a current sub-time period further comprises:
determining, for each display position, a current click-through rate threshold and a current value of each of the plurality of pieces of the first-type display information based at least in part on a target number of displaying each piece of the first-type display information in the current sub-time period; and
determining the target display information to be allocated to the display position based on the current click-through rate threshold and the current value of each of the plurality of pieces of the first-type display information, the estimated click-through rate of the display position, and the value of each of the plurality of pieces of the second-type display information.

4. The method according to claim 3 , wherein the current click-through rate threshold of each of the plurality of pieces of the first-type display information in the current sub-time period is determined by:
dividing, for each display position, the current sub-time period equally into u sub-time intervals;
determining, for each of the plurality of pieces of the first-type display information, a click-through rate threshold of each piece of the first-type display information in each of u sub-time intervals using following equations:

P
i

c
⁢
t
⁢
r

←

n
i

⁢

r
⁡

(

t
j
l

)

+

ɛ

c
⁢
t
⁢
r

n
i
′

⁡

(

t
j
l

)

+

ɛ

c
⁢
t
⁢
r

I
i
ctr

←

clamp
⁢

Assignees

• Beijing Bytedance Network Tech Co Ltd

Inventors

• Xiao Shenke
• Tang Pingzhong
• Zhang Lihua
• Yang Xiwang
• Gu Changsheng

Classifications

• G06Q10/40 Primary

  Business processes related to social networking or social networking services · CPC title

• G06F3/14

  Digital output to display device {; Cooperation and interconnection of the display device with other functional units} · CPC title

• G06F40/106 Primary

  Display of layout of documents; Previewing · CPC title

• G06Q30/02

  Marketing; Price estimation or determination; Fundraising · CPC title

• G06N20/00

  Machine learning · CPC title