WO2023207267A1

WO2023207267A1 - Data transmission method, model training method, device, and computer-readable medium

Info

Publication number: WO2023207267A1
Application number: PCT/CN2023/076528
Authority: WO
Inventors: 李彦志
Original assignee: 中兴通讯股份有限公司
Priority date: 2022-04-28
Filing date: 2023-02-16
Publication date: 2023-11-02
Also published as: CN117015014A

Abstract

Embodiments of the present application provide a data transmission method, a model training method, an electronic device, and a computer-readable medium. The data transmission method comprises: under the condition that the screen-off of a terminal device is detected, predicting an estimated duration of the present screen-off; determining, according to the current network state parameter of the terminal device, a target duration required for performing data transmission this time; and determining, according to the estimated duration and the target duration, whether to perform data transmission.

Description

Data transmission method, model training method, equipment, computer-readable media

Technical field

Embodiments of the present application relate to the field of communication technology, and in particular to data transmission methods, model training methods, electronic devices, and computer-readable media.

Background technique

Many terminal devices now automatically back up certain important data on the device to the server in the background on a regular basis. In the future era of the Internet of Everything, in addition to data transmission between terminal devices and servers, the frequency of data transmission between terminal devices will be higher and the amount of data will be larger. During data transmission, the terminal device will occupy the network bandwidth as much as possible; it will also quickly read and write, compress and decompress the data. These processing processes have a lot of overhead on the central processing unit (CPU, Central Processing Unit) and memory. Some data transmitted by terminal devices do not have high timeliness requirements, such as system upgrades and downloading resource packages, regular backup of local data, etc. These can all use background automatic transmission strategies. If the user is also using other applications on the terminal device during the background data transmission of the terminal device, the background data transmission will compete with the applications used by the user for network, CPU, memory and other resources, resulting in the user being very likely to encounter network lag and open the Situations such as long application time and slow reading of data will greatly reduce the user experience.

Contents of the invention

Embodiments of the present application provide a data transmission method, a model training method, an electronic device, and a computer-readable medium.

Embodiments of the present application provide a data transmission method, which includes: when a terminal device is detected to be turned off, predicting the estimated duration of the screen off; and determining the current data transmission based on the current network status parameters of the terminal device. The required target duration; determine whether to perform data transmission based on the estimated duration and the target duration.

Embodiments of the present application provide a model training method, including: receiving historical screen information data within a preset historical time period sent by a terminal device; wherein the historical screen information data includes: historical screen screen time, the next historical screen time screen time, historical screen mode, and historical date category; perform model training based on the historical screen data to obtain a prediction model corresponding to the historical date category; and send the prediction model corresponding to the historical date category to the terminal device.

An embodiment of the present application provides an electronic device, including: at least one processor; and a memory. At least one program is stored on the memory. When the at least one program is executed by the at least one processor, any of the above data transmissions are implemented. method, or any of the above model training methods.

Embodiments of the present application provide a computer-readable medium. A computer program is stored on the computer-readable medium. When the computer program is executed by a processor, any of the above-mentioned data transmission methods or any of the above-mentioned model training methods are implemented.

Description of the drawings

Figure 1 is a flow chart of a data transmission method provided by an embodiment of the present application;

Figure 2 is a flow chart of a model training method provided by another embodiment of the present application;

Figure 3 is a block diagram of a data transmission device provided by another embodiment of the present application;

Figure 4 is a block diagram of a model training device provided by another embodiment of the present application.

Detailed ways

In order to enable those skilled in the art to better understand the technical solutions of the present application, the data transmission method, model training method, electronic device, and computer-readable medium provided by the embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

Example embodiments will be described more fully below with reference to the accompanying drawings, which may, however, be embodied in different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully enable those skilled in the art to fully understand the scope of the disclosure.

The embodiments of the present application and the features in the embodiments can be combined with each other unless there is any conflict.

As used herein, the term "and/or" includes any and all combinations of at least one associated listed item.

The terminology used herein is used to describe particular embodiments only and is not intended to be limiting of the application. As used herein, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms "comprising" and/or "made of" are used in this specification, the presence of stated features, integers, steps, operations, elements and/or components is specified but does not exclude the presence or Add at least one other feature, integer, step, operation, element, component and/or group thereof.

Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be construed to have meanings consistent with their meanings in the context of the relevant art and the present application, and will not be construed as having idealized or excessively formal meanings, Unless expressly so limited herein.

Figure 1 is a flow chart of a data transmission method provided by an embodiment of the present application.

Referring to Figure 1, one embodiment of the present application provides a data transmission method, which can be applied to terminal devices. The method includes:

Step 100: When it is detected that the screen of the terminal device is off, predict the estimated duration of the screen off this time.

In some exemplary embodiments, the estimated duration of the current pause screen can be predicted using methods well known to those skilled in the art.

In some exemplary embodiments, predicting the estimated duration of the current screen pause period includes: obtaining the screen pause moment and screen pause mode of the terminal device; and predicting the estimated duration based on the screen pause moment and screen pause mode.

In some exemplary implementations, historical breath-holding moments, breath-holding modes, and their corresponding breath-holding durations can be combined to predict the current breath-holding time and breath-holding mode based on the current breath-holding time and breath-holding mode. Estimate the duration.

In other exemplary embodiments, predicting the estimated duration based on the screen-breathing moment and the screen-breathing mode includes: determining the date category corresponding to the screen-blanking moment and the time period in which the screen-blanking moment is located; The time period in which the screen pause moment is located is input into the prediction model corresponding to the date category to obtain the estimated duration.

In some exemplary embodiments, the screen-off moment refers to the moment when the screen of the terminal device turns off.

In some exemplary embodiments, the screen rest mode is divided into a manual screen rest mode and an automatic screen rest mode. Since the user's status is different in these two modes, specifically, the manual screen rest mode indicates that the user may have temporarily completed the screen rest mode. A certain operation on the terminal device may require temporary screen suspension to avoid external unexpected factors. Most of the automatic screen suspension methods are used by users who want the screen to be on. This may be to wait for the countdown to buy goods, or to copy a paragraph on the terminal device. Text, no screen clicks for a long time. Different resting modes have an impact on the model training results and prediction results, so it is necessary to add the resting mode as a feature for prediction.

For example, the above-mentioned screen-resting mode may be an element in the screen-resting mode set S={s1, s2}, s1 represents the manual screen-resting mode, and s2 represents the automatic screen-closing mode.

In some exemplary embodiments, the date category corresponding to the screen-stop moment refers to the date to which the date of the screen-stop moment belongs. category. For example, for the date category set D={d1, d2,..., dn1}, d1 represents working days, d2 represents holidays, d3 represents shopping days, and so on. Then if the date corresponding to the screen off moment is a working day, the date category corresponding to the screen off moment is considered to be a working day, and so on.

In some exemplary embodiments, a day is divided into Y segments according to predetermined standards, such as each segment is uniformly X minutes (such as 5 minutes), then the day can be divided into 1440/X=Y segments, expressed as a set of time segments Y={ y1, y2,...,ym}, y1 represents the first time period, y2 represents the second time period, and so on; or, divide according to the time period, such as 00:00-6:00 according to each period X1 Minutes are divided, and 6:00-9:00 is divided into X2 minutes per segment, and so on, to get Y segments.

In this way, it is possible to calculate which time period corresponds to the screen-off moment, that is, the time period in which the screen-off moment is located.

In some exemplary embodiments, since users have different habits of using terminal devices on dates corresponding to different date categories, when training the model, it is necessary to train different prediction models for different date categories, that is, different date categories. The corresponding prediction models are different.

The embodiment of this application does not limit the specific classification method of date categories. For example, date categories can be categories such as working days, holidays, shopping days, etc.

Step 101: Determine the target duration required for this data transmission according to the current network status parameters of the terminal device.

In some exemplary embodiments, the network status parameter is the network transmission rate; determining the target duration required for this data transmission according to the current network status parameter of the terminal device may include: determining the target duration as the data amount for this data transmission and The ratio of network transmission rates.

Step 102: Determine whether to perform data transmission based on the estimated duration and target duration.

In some exemplary embodiments, determining whether to perform data transmission based on the estimated duration and the target duration includes at least one of the following: when the difference between the estimated duration and the target duration is greater than or equal to a preset threshold, determine to proceed Data transmission; when the difference between the estimated duration and the target duration is less than the preset threshold, it is determined that no data transmission will be performed.

In some exemplary embodiments, the preset threshold may be a positive value or a negative value. If the preset threshold is a negative value, the absolute value of the preset threshold should be less than a certain value, which means that the time required for data transmission this time is slightly longer than the screen time, and the data transmission is considered possible. of.

In some exemplary embodiments, the preset threshold can be set arbitrarily by engineers according to actual conditions.

In some exemplary embodiments, the method further includes: sending the historical screen-off data of the terminal device within a preset historical time period to the server; wherein the historical screen-off data includes: the historical screen-off moment and the next historical screen-on time. Time, historical screen mode, historical date category; receiving the prediction model corresponding to the historical date category sent by the server through model training based on historical screen data.

In some exemplary embodiments, in order to ensure that the prediction model better fits the user's latest habits, the terminal device will continue to obtain the screen-time data and send it to the server. The server can clear the historical screen-time data and use the new screen-time data for model training. , so that the weights of the model parameters can be adjusted accordingly as the user's screen habit changes, ensuring that the prediction model can dynamically fit the user's habits.

In the embodiment of the present application, any interest screen data that does not correspond to the current interest screen can be considered as historical interest screen data, that is, the interest screen data corresponding to the interest screen that has already occurred.

In some exemplary embodiments, the terminal device may send the breath-taking data corresponding to the breath-taking screen to the server each time it detects the breath-taking screen, or may obtain all historical breath-taking data within the preset historical time period. , and then send all historical screen data within the preset historical time period to the server.

The data transmission method provided by the embodiment of the present application predicts the estimated duration of the current screen time, and determines whether to perform data transmission based on the estimated duration of the current screen time and the target duration required for this data transmission. Compared with the traditional The method of waking up the target application to start running when the network status, CPU status, memory status, charging status, etc. of the terminal device are all lower than the set threshold. The data transmission method in the embodiment of the present application only needs to be based on the estimation of the current screen. The duration and the target duration required for this data transmission determine whether to perform data transmission. It only needs to ensure that the user's use will not be affected during the data transmission. Therefore, the data transmission method of the embodiment of the present application is more flexible, practical, and applicable There are more occasions, thus improving the user experience.

Figure 2 is a flow chart of a model training method provided by another embodiment of the present application.

Referring to Figure 2, another embodiment of the present application provides a model training method, which can be applied to a server. The method can include:

Step 200: Receive historical screen-holding data within a preset historical time period sent by the terminal device; wherein the historical screen-holding data includes: historical screen-holding time, next historical screen-on time, historical screen-holding mode, and historical date category.

In this embodiment of the present application, as long as the screen data does not correspond to the current screen of the terminal device, it can be considered as historical screen data, that is, the screen data corresponding to the screen that has already occurred.

In some exemplary embodiments, the historical time period can be set arbitrarily in advance, for example, set to 60 days.

In some exemplary embodiments, the historical screen data within the preset historical time period refers to the historical screen data corresponding to the screen that occurred within the preset historical time period.

In some exemplary embodiments, the historical screen-off moment refers to the moment when the screen of the terminal device turns off.

In some exemplary embodiments, the next historical moment when the screen is turned on refers to the time when the terminal device turns on the screen again after the screen is turned off.

In some exemplary embodiments, the historical screen resting mode is divided into a manual screen resting mode and an automatic screen resting mode. Since the status of the user in these two modes is different, specifically, the manual screen resting mode indicates that the user may have temporarily completed the screen resting mode. An operation on the terminal device is interrupted, or the screen needs to be temporarily paused to avoid external emergencies. Most of the automatic screen pause methods are used by users who want the screen to be on. This may be to wait for the countdown to buy goods, or to copy the messages on the terminal device. A piece of text that has not been clicked on the screen for a long time. Different resting modes have an impact on the model training results and prediction results, so it is necessary to add the resting mode as a feature for prediction.

For example, the above historical screen rest mode may be an element in the screen rest mode set S={s1, s2}, where s1 represents the manual screen rest mode and s2 represents the automatic screen rest mode.

In some exemplary embodiments, the historical date category corresponding to the historical screen moment refers to the date category to which the date of the historical screen moment belongs. For example, for the date category set D={d1, d2,..., dn}, d1 represents working days, d2 represents holidays, d3 represents shopping days, and so on. Then if the date corresponding to the historical screen time is a working day, the historical date category corresponding to the historical screen time is considered to be a working day, and so on.

Step 201: Perform model training based on historical screen data to obtain a prediction model corresponding to the historical date category.

In some exemplary embodiments, the historical screen data of different terminal devices should be trained separately. That is to say, the prediction models corresponding to different terminal devices are different.

In some exemplary embodiments, model training is performed based on historical screen data to obtain predictions corresponding to historical date categories. The model includes: dividing historical screen data according to historical date categories; performing model training based on historical screen data of the same historical date category to obtain a prediction model corresponding to the historical date category.

In some exemplary embodiments, performing model training based on historical screen data of the same historical date category to obtain a prediction model corresponding to the historical date category includes: for historical screen data of the same historical date category, separately determining each piece of historical screen data Corresponding training samples; among which, the training samples include: historical screen-off moment, next historical screen-on moment, historical screen-off mode, screen-off duration type, and the time period of the historical screen-off moment; according to all historical date categories The training samples corresponding to the historical information screen data are used for model training to obtain the prediction model corresponding to the historical date category.

In this way, it is possible to calculate which time period the historical pause moment corresponds to, that is, the time period in which the historical pause moment is located.

In some exemplary embodiments, the screen-off duration refers to the time difference between the next historical screen-on time and the historical screen-off time.

In some exemplary embodiments, historical screen data is classified according to screen duration. Historical screen data with a screen duration less than A0 minutes are marked as n0, and historical screen data with a screen duration between A0 and A1 minutes are marked. The data is marked as n1, the historical screen data with a screen time between A1 and A2 minutes is marked as n2, the historical screen data with a screen time between A2 and A3 minutes is marked as n3, and so on. In this way, the screen duration type can be considered as an element in the screen duration type set N = {n0, n1, n2,...}, n0 means that the screen time is less than A0 minutes, n1 means that the screen time is between A0-A1 minutes. time, and so on.

In some exemplary embodiments, during the model training process, the historical screen time, historical screen mode, and the time period of the historical screen time in the training samples are used as the input of the prediction model, and the next time in the training sample is The difference between the historical screen-on time and the historical screen-off time is used as the output of the prediction model for model training. During the model training process, the screen-off duration type will be used for related calculations.

In some exemplary embodiments, model training may be the training of classic machine learning algorithms such as decision tree (DT, Decision Tree), logistic regression (LR, Logistic Regression), support vector machine (SVM, Support Vector Machine), etc. It can be used for neural network models such as fully connected neural networks, convolutional neural networks (CNN, Convolutional Neural Networks), recurrent neural networks (RNN, Recurrent Neural Networks), long short-term memory networks (LSTM, Long Short-Term Memory), etc. Training of a combination of one or more models.

Step 202: Send the prediction model corresponding to the historical date category to the terminal device.

In order to better present the model training method and data transmission method of the embodiment of the present application, the model training process and the data transmission process are described in detail below through a specific example. The examples listed are not used to limit the protection scope of the embodiment of the present application.

Example

This example describes the model training process and data transmission process, including:

1. Each time the terminal device detects the interest screen, it records the interest screen data corresponding to the interest screen, and sends the interest screen data corresponding to the interest screen to the server.

In this step, the screen rest data includes the screen rest time t1, the next screen on time t2, the screen rest mode si, and the date category dj. i and j are integers greater than or equal to 1. In other words, the information screen data can be represented by a four-tuple (dj, t1, t2, si).

Among them, the screen-off time t1 refers to the time when the screen of the terminal device goes out.

The next time the screen is turned on t2 refers to the time when the terminal device turns on the screen again after the screen is turned off.

Among them, the screen rest mode si is an element in the screen rest mode set S = {s1, s2}, s1 represents the manual screen rest mode, and s2 represents the automatic screen rest mode.

Among them, the date category dj is an element in the date category set D = {d1, d2,..., dn1}, d1 represents a working day, d2 represents a holiday, d3 represents a shopping day, and so on.

In this step, if the screen-holding behavior of the terminal device has ended, the screen-holding data corresponding to the screen-holding time can be considered to be historical screen-holding data.

2. The server receives the historical screen data of the terminal device within the preset historical time period, divides the historical screen data according to historical date categories, and determines each piece of historical screen data for the historical screen data of the same historical date category. Training samples corresponding to the data; perform model training based on the training samples corresponding to all historical screen data of the same historical date category to obtain a prediction model corresponding to the historical date category.

In this step, the training samples corresponding to the historical screen-disabled data include: historical screen-disabled time t1, the next historical screen-disabled moment t2, historical screen-disabled mode si, screen-disabled duration type nk, and the time period of the historical screen-disabled moment. yp.

Among them, the historical screen data is classified according to the screen time. The historical screen data with a screen time of less than A0 minutes is marked as n0. The historical screen data with a screen time between A0 and A1 minutes is marked as n1. The historical screen data with a screen duration between A1 and A2 minutes is marked as n2, the historical screen data with a screen duration between A2 and A3 minutes is marked as n3, and so on. In this way, the screen time type nk can be considered as an element in the screen time type set N = {n0, n1, n2,...}, n0 means that the screen time is less than A0 minutes, n1 means that the screen time is between A0 and A1 minutes. in between, and so on.

Among them, the day is divided into Y segments according to predetermined standards, for example, each segment is unified to X minutes (such as 5 minutes), then the day can be divided into 1440/X=Y segments, expressed as a time segment set Y={y1, y2,... , ym}, y1 represents the first time period, y2 represents the second time period, and so on; or, divide according to the time period, for example, 00:00-6:00 is divided according to X1 minutes per period, 6: 00-9:00 is divided into X2 minutes per segment, and so on, to get Y segment.

In this way, it is possible to calculate which time period the historical pause moment corresponds to, that is, the time period in which the historical pause moment is located. In other words, the time period yp where the historical screen moment is located can be considered as the elements in the time period set Y = {y1, y2,..., ym}.

3. The server sends the prediction model corresponding to the historical date category to the terminal device.

4. When the terminal device detects that the screen is off, the terminal device obtains the screen off time and screen off mode of the terminal device.

5. The terminal device determines the date category corresponding to the screen-off moment and the time period of the screen-off moment.

6. The terminal device inputs the screen pause moment, screen pause mode, and the time period of the screen pause moment into the prediction model corresponding to the date category to obtain the estimated duration of the current screen pause.

7. The terminal device determines that the target duration required for this data transmission is the data volume and network for this data transmission. The ratio of transmission rates.

8. When the difference between the estimated duration and the target duration is greater than or equal to the preset threshold, the terminal device determines to transmit data; when the difference between the estimated duration and the target duration is less than the preset threshold , the terminal device determines not to transmit data.

Another embodiment of the present application provides an electronic device, including: at least one processor; and a memory. At least one program is stored on the memory. When at least one program is executed by at least one processor, any one of the above data transmission methods is implemented. Or any of the above model training methods.

Among them, the processor is a device with data processing capabilities, including but not limited to a central processing unit (CPU), etc.; the memory is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically such as SDRAM). , DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory (FLASH).

In some embodiments, the processor and memory are connected to each other through a bus, and then to other components of the computing device.

Another embodiment of the present application provides a computer-readable medium. A computer program is stored on the computer-readable medium. When the computer program is executed by a processor, any of the above-mentioned data transmission methods or any of the above-mentioned model training methods are implemented.

Figure 3 is a block diagram of a data transmission device provided by another embodiment of the present application.

Referring to Figure 3, another embodiment of the present application provides a data transmission method, including: a prediction module 301, configured to predict the estimated duration of the current screen pause when the terminal device is detected to pause the screen; a determination module 302, It is set to determine the target duration required for this data transmission based on the current network status parameters of the terminal device; and determine whether to perform data transmission based on the estimated duration and the target duration.

In some exemplary embodiments, the prediction module 301 is specifically configured to: when detecting that the terminal device is screen-stopping, obtain the screen-off time and screen-off mode of the terminal device; Screen mode predicts the estimated duration.

In some exemplary embodiments, the prediction module 301 is specifically configured to implement the prediction of the estimated duration according to the screen-blanking moment and the screen-blanking mode in the following manner: determine the date category corresponding to the screen-blanking moment, and the time period in which the screen-blanking moment is located; input the screen-blanking moment, the screen-blanking mode and the time period in which the screen-blanking moment is located into the prediction model corresponding to the date category to obtain the estimated duration .

In some exemplary embodiments, it also includes: a first model training module 303, configured to send historical breath-screen data of the terminal device within a preset historical time period to the server; wherein the historical breath-screen data includes : historical screen-off moment, next historical screen-on moment, historical screen-off mode, and historical date category; receiving a prediction model corresponding to the historical date category sent by the server and obtained by model training based on the historical screen-off data.

In some exemplary embodiments, the network status parameter is a network transmission rate; the determination module 302 is specifically configured to implement the determination of the target duration according to the current network status parameter of the terminal device in the following manner: determine the target duration as this time The ratio of the amount of data transmitted to the network transmission rate.

In some exemplary embodiments, the determination module 302 is specifically configured to use at least one of the following methods to determine whether to perform data transmission based on the estimated duration and the target duration: When the difference between the duration is greater than or equal to the preset threshold, it is determined to perform data transmission; when the difference between the estimated duration and the target duration is less than the preset threshold, it is determined not to perform data transmission .

The specific implementation process of the above data transmission device is the same as the specific implementation process of the data transmission method in the previous embodiment, and will not be described again here.

Referring to Figure 4, another embodiment of the present application provides a model training method, including: a receiving module 401, configured to receive historical information screen data within a preset historical time period sent by a terminal device; wherein, the historical information screen data The data includes: historical screen-off moment, next historical screen-on moment, historical screen-off mode, and historical date category; the second model training module 402 is configured to perform model training based on the historical screen-off data to obtain the historical date category correspondence. The prediction model; the sending module 403 is configured to send the prediction model corresponding to the historical date category to the terminal device.

In some exemplary embodiments, the second model training module 402 is specifically configured to: divide the historical screen data according to the historical date category; perform model training based on the historical screen data of the same historical date category. The prediction model corresponding to the historical date category.

In some exemplary embodiments, the second model training module 402 is specifically configured to perform model training based on historical screen data of the same historical date category to obtain a prediction model corresponding to the historical date category in the following manner: for For historical screen data of the same historical date category, a training sample corresponding to each piece of historical screen data is determined respectively; wherein, the training samples include: the historical screen screen moment and the next historical screen light time. , the historical screen mode, screen duration type, and the time period in which the historical screen moment is located; perform model training based on the training samples corresponding to all the historical screen data of the same historical date category to obtain the above Prediction model corresponding to historical date category.

The specific implementation process of the above model training device is the same as the specific implementation process of the model training method in the previous embodiment, and will not be described again here.

Those of ordinary skill in the art can understand that all or some steps, systems, and functional modules/units in the devices disclosed above can be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may consist of several physical components. Components execute cooperatively. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit . Such software may be distributed on computer-readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). As is known to those of ordinary skill in the art, the term computer storage media includes volatile and nonvolatile media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. removable, removable and non-removable media. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disk (DVD) or other optical disk storage, magnetic cassettes, tapes, disk storage or other magnetic storage, or may be used Any other medium that stores the desired information and can be accessed by a computer. Additionally, it is known to those of ordinary skill in the art that communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media .

Example embodiments have been disclosed herein, and although specific terms are employed, they are used and should be interpreted in a general illustrative sense only and not for purpose of limitation. In some instances, it will be apparent to those skilled in the art that features, characteristics and/or elements described in connection with a particular embodiment may be used alone, or may be used in conjunction with other embodiments, unless expressly stated otherwise. Features and/or components used in combination. Accordingly, it will be understood by those skilled in the art that various changes in form and details may be made without departing from the scope of the application as set forth in the appended claims.

Claims

A data transmission method including:

When the terminal device's screen is detected to be off, predict the estimated duration of the screen off;

Determine the target duration required for this data transmission according to the current network status parameters of the terminal device;

Determine whether to perform data transmission based on the estimated duration and the target duration.
The data transmission method according to claim 1, wherein the predicting the estimated duration of this pause screen includes:

Obtain the screen pause time and screen pause mode of the terminal device;

The estimated duration is predicted based on the screen-off moment and the screen-off mode.
The data transmission method according to claim 2, wherein predicting the estimated duration based on the screen-off moment and the screen-off mode includes:

Determine the date category corresponding to the screen-off moment and the time period in which the screen-off moment is located;

The estimated duration is obtained by inputting the screen pause moment, the screen pause mode, and the time period in which the screen pause moment is located into the prediction model corresponding to the date category.
The data transmission method according to claim 3, further comprising:

Send the historical screen-disabled data of the terminal device within the preset historical time period to the server; wherein the historical screen-disabled data includes: historical screen-disabled moment, the next historical screen-disabled moment, historical screen-disabled mode, and historical date. category;

Receive the prediction model corresponding to the historical date category and obtained by model training based on the historical screen data and sent by the server.
The data transmission method according to any one of claims 1-4, wherein the network status parameter is a network transmission rate;

Determining the target duration required for this data transmission based on the current network status parameters of the terminal device includes:

The target duration is determined as the ratio of the amount of data to be transmitted this time and the network transmission rate.
The data transmission method according to any one of claims 1 to 4, wherein determining whether to perform data transmission based on the estimated duration and the target duration includes at least one of the following:

If the difference between the estimated duration and the target duration is greater than or equal to a preset threshold, determine to perform data transmission;

If the difference between the estimated duration and the target duration is less than a preset threshold, it is determined not to perform data transmission.
A model training method including:

Receive historical screen pause data sent by the terminal device within a preset historical time period; wherein the historical screen pause data includes: historical screen pause moment, next historical screen turn-on moment, historical screen pause mode, and historical date category;

Perform model training based on the historical screen data to obtain a prediction model corresponding to the historical date category;

Send the prediction model corresponding to the historical date category to the terminal device.
The model training method according to claim 7, wherein said performing model training based on the historical screen data to obtain the prediction model corresponding to the historical date category includes:

Divide the historical screen data according to the historical date categories;

Model training is performed based on historical screen data of the same historical date category to obtain a prediction model corresponding to the historical date category.
The model training method according to claim 8, wherein said performing model training based on historical screen data of the same historical date category to obtain a prediction model corresponding to the historical date category includes:

For the historical screen data of the same historical date category, a training sample corresponding to each piece of historical screen data is determined respectively; wherein, the training samples include: the historical screen time moment, the next historical screen turn-on time Time, the historical screen mode, the type of screen duration, and the time period in which the historical screen moment is located;

Model training is performed based on the training samples corresponding to all the historical screen data of the same historical date category to obtain a prediction model corresponding to the historical date category.
An electronic device including:

at least one processor;

Memory, at least one program is stored on the memory, and when the at least one program is executed by the at least one processor, the data transmission method according to any one of claims 1-6 is implemented, or claims 7-9 Any of the model training methods described above.
A computer-readable medium with a computer program stored on the computer-readable medium. When the computer program is executed by a processor, the data transmission method of any one of claims 1-6 is implemented, or claims 7-9 Any of the model training methods described above.