WO2019062409A1 - 后台应用程序管控方法、存储介质及电子设备 - Google Patents

后台应用程序管控方法、存储介质及电子设备 Download PDF

Info

Publication number
WO2019062409A1
WO2019062409A1 PCT/CN2018/102187 CN2018102187W WO2019062409A1 WO 2019062409 A1 WO2019062409 A1 WO 2019062409A1 CN 2018102187 W CN2018102187 W CN 2018102187W WO 2019062409 A1 WO2019062409 A1 WO 2019062409A1
Authority
WO
WIPO (PCT)
Prior art keywords
background application
preset
training
reference model
prediction results
Prior art date
Application number
PCT/CN2018/102187
Other languages
English (en)
French (fr)
Inventor
曾元清
Original Assignee
Oppo广东移动通信有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Oppo广东移动通信有限公司 filed Critical Oppo广东移动通信有限公司
Publication of WO2019062409A1 publication Critical patent/WO2019062409A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
    • G06F11/302—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system component is a software system
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3051—Monitoring arrangements for monitoring the configuration of the computing system or of the computing system component, e.g. monitoring the presence of processing resources, peripherals, I/O links, software programs
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
    • G06F11/3452—Performance evaluation by statistical analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5011—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals
    • G06F9/5022—Mechanisms to release resources

Definitions

  • the present application belongs to the field of communications technologies, and in particular, to a background application control method, a storage medium, and an electronic device.
  • the application provides a background application control method, a storage medium, and an electronic device, which can improve the accuracy of controlling the application.
  • the embodiment of the present application provides a background application control method, which is applied to an electronic device, and the method includes:
  • an embodiment of the present application provides a storage medium on which a computer program is stored, and when the computer program runs on a computer, causes the computer to execute the background application control method described above.
  • an embodiment of the present application provides an electronic device, including a processor and a memory, where the memory has a computer program, and the processor calls the computer program, wherein the processor further performs:
  • FIG. 1 is a schematic system diagram of a background application control device according to an embodiment of the present application.
  • FIG. 2 is a schematic diagram of an application scenario of a background application control device according to an embodiment of the present disclosure
  • FIG. 3 is a schematic flowchart of a background application control method according to an embodiment of the present application.
  • FIG. 4 is another schematic flowchart of a background application control method according to an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram of a background application control device according to an embodiment of the present application.
  • FIG. 6 is another schematic structural diagram of a background application control device according to an embodiment of the present application.
  • FIG. 7 is a schematic diagram of another application scenario of a background application control device according to an embodiment of the present disclosure.
  • FIG. 8 is still another schematic structural diagram of a background application control device according to an embodiment of the present application.
  • FIG. 9 is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure.
  • FIG. 10 is another schematic structural diagram of an electronic device according to an embodiment of the present application.
  • the embodiment of the present application provides a background application control method, which is applied to an electronic device, and the method includes:
  • the reference model is a cyclic neural network model.
  • the cyclic neural network model includes a network subunit and a classifier
  • the step of training the first binary vector as the training data into the reference model, and obtaining the optimized parameters of the reference model after the training includes:
  • Training is performed according to the loss value to obtain the optimization parameter.
  • the probability that the intermediate value is input to the classifier to obtain a plurality of the prediction results includes:
  • Z K is the target intermediate value
  • C is the number of categories of prediction results
  • Z j is the jth intermediate value
  • the obtaining the loss value according to the plurality of the prediction results and the plurality of the probabilities corresponding thereto specifically includes:
  • the obtaining the loss value according to the plurality of the prediction results and the plurality of the probabilities corresponding thereto specifically includes:
  • E is the average.
  • the obtaining the loss value according to the plurality of the prediction results and the plurality of the probabilities corresponding thereto specifically includes:
  • a loss value is obtained based on a plurality of the prediction results and a plurality of the probabilities corresponding thereto.
  • the training according to the loss value specifically includes:
  • Training is performed using a stochastic gradient descent method based on the loss value.
  • the method further includes: before the acquiring a plurality of consecutive use states of the preset background application in the preset time period, the method further includes:
  • a background application that scores above a preset rating threshold is set as the default background application.
  • the scoring the plurality of background applications according to the running data of the background application includes:
  • the plurality of background applications are scored according to the running data of the background application based on the third preset formula, wherein the third preset formula is:
  • memory i is the memory usage ratio
  • couter i is the total usage time
  • t i is the total usage time
  • seq i is the latest usage order of the default background application
  • Group is the grouping coefficient
  • G f is the sinking factor
  • T f is Time weakening factor.
  • the controlling the preset background application according to the prediction result specifically includes:
  • the preset background application is retained when the predicted result is reserved.
  • references to "an embodiment” herein mean that a particular feature, structure, or characteristic described in connection with the embodiments can be included in at least one embodiment of the present application.
  • the appearances of the phrases in various places in the specification are not necessarily referring to the same embodiments, and are not exclusive or alternative embodiments that are mutually exclusive. Those skilled in the art will understand and implicitly understand that the embodiments described herein can be combined with other embodiments.
  • FIG. 1 is a schematic diagram of a system of a background application control device according to an embodiment of the present application.
  • the background application control device is mainly configured to: obtain a plurality of consecutive use states of the preset background application in a preset time period, and form the obtained multiple continuous use states into a first binary value vector; acquire a reference model, and The first binary vector is used as the training data input reference model for training, and the optimized parameters of the reference model after training are obtained; and the plurality of consecutive use states of the preset background application corresponding to the current time are obtained, and multiple times corresponding to the current time are obtained.
  • the continuous use state forms a second binary vector, performs prediction according to the reference model, the optimization parameter, and the second binary vector, generates a prediction result, and then determines, according to the prediction result, whether the preset background application needs to be used, to preset
  • the background application is managed, such as shutting down, or freezing.
  • FIG. 2 is a schematic diagram of an application scenario of a background application control device according to an embodiment of the present application.
  • the background application management device detects that the application running in the background of the electronic device includes a preset background application a, a preset background application b, and a preset background application c; Presetting the reference model A corresponding to the background application a, the reference model B corresponding to the preset background application b, and the reference model C corresponding to the preset background application c; whether the preset background application a needs to be The probability of use is predicted to obtain the probability a', and the probability of whether the preset background application b needs to be used is predicted by the reference model B, and the probability b' is obtained, and the reference model C needs to be used for the preset background application c.
  • the probability is predicted to obtain the probability c′; according to the probability a′, the probability b′ and the probability c′, the preset background application a running in the background, the preset background application b and the preset background application c are controlled, for example The default background application b with the lowest probability is turned off.
  • the embodiment of the present application provides a background application control method, and the execution entity of the background application control method may be a background application control device provided by the embodiment of the present application, or an electronic device integrated with the background application control device, wherein
  • the background application management device can be implemented in hardware or software.
  • the background application management method includes: acquiring a plurality of consecutive use states of the preset background application in a preset time period, and forming the obtained plurality of continuous use states into a first binary vector; acquiring the reference model, and A binary vector is trained as a training data input reference model to obtain optimized parameters of the reference model after training; and a plurality of consecutive use states corresponding to the current time are obtained by the preset background application, and multiple consecutive uses corresponding to the current time are obtained.
  • the state forms a second binary vector, predicts according to the reference model, the optimization parameter and the second binary vector, generates a prediction result, and controls the preset background application according to the prediction result.
  • FIG. 3 is a schematic flowchart of a background application control method according to an embodiment of the present application.
  • the background application management and control method provided by the embodiment of the present application is applied to an electronic device, and the specific process may be as follows:
  • the preset time period can be set according to the usage period, such as 8:00 to 12:00 in the morning, 6 to 10 in the evening, or Monday to Friday, and Saturday to Sunday.
  • the preset time period can be several hours or days, or even longer.
  • a plurality of continuous use states of the preset background application are obtained in the preset time period, and the use state is whether the preset background application is in use, if the use is recorded as 1, and if not used, the record is 0.
  • the usage status is continuous. If the usage status is recorded every 5 minutes within one hour, 12-13 consecutive use statuses can be obtained.
  • the judgment criterion of the usage status can be as long as the user can call once within 5 minutes. I think it is in use. Of course, the interval can be set to other values, such as 10 minutes, 3 minutes, and so on.
  • the default background application can be a web browsing application, a video playing application, a news browsing application, and the like.
  • the data of the first binary vector is a set of binary data, such as ⁇ 0, 1, 1, 0, 0, 0, 1 ⁇ , wherein the data 0 indicates that the preset background application is not called during the time period, and the data 1 Indicates that the default background application is called at least once during this time period.
  • the reference model can be a cyclic neural network model.
  • the reference model can be a hybrid neural network model, a Gaussian mixture model or a convolutional neural network model.
  • the cyclic neural network model includes network sub-units and classifiers. Specifically, the first binary vector is used as the training data input network subunit to obtain an intermediate value; the intermediate value is input to the classifier to obtain a probability corresponding to the plurality of prediction results; and the loss value is obtained according to the plurality of prediction results and the plurality of probabilities corresponding thereto Training based on the loss value to obtain optimized parameters.
  • the network subunit specifically includes an input layer, a hidden layer, and an output layer.
  • the first binary vector is used as the input layer of the training data input network subunit, the input layer inputs the first binary vector into the hidden layer, and the hidden layer cyclically calculates the intermediate value of the first binary vector, and then passes the intermediate value through the output layer. Output.
  • the intermediate value may be input to the classifier based on the first preset formula to obtain a probability corresponding to the plurality of prediction results, where the first preset formula is:
  • Z K is the target intermediate value
  • C is the number of categories of prediction results
  • Z j is the jth intermediate value.
  • the median value of the target is the median value of the kth.
  • the prediction result can be set as needed. For example, if the background application's prediction result includes cleaning and not cleaning, the number of categories C of the prediction result is 2.
  • the loss value may be based on a plurality of prediction results and a plurality of probabilities corresponding thereto according to a fourth preset formula, wherein the fourth preset formula is:
  • the loss value may be obtained according to the plurality of prediction results and the plurality of probabilities corresponding thereto according to the second preset formula, where the second preset formula is:
  • C is the number of categories of prediction results
  • y k is the true value
  • the true value can be 0 or 1.
  • the result of the first preset formula. E is the average value, and the loss value obtained by the fourth preset formula can be obtained, and then the average value is obtained by averaging, and the data is more accurate.
  • the acquisition loss value may be acquired according to the cross entropy or the like in addition to the above method.
  • the training can be performed using a stochastic gradient descent method based on the loss value. Training can also be performed according to the batch gradient descent method or the gradient descent method.
  • Training is performed using the stochastic gradient descent method, and the training can be completed when the loss value is equal to or less than the preset loss value. It is also possible to complete the training when there are no changes in the two or more loss values continuously acquired. Of course, it is also possible to directly set the number of iterations of the random gradient descent method according to the loss value. After the number of iterations is completed, the training is completed. After the training is completed, each parameter of the reference model at this time is obtained, and the each parameter is saved as an optimization parameter, and when the prediction is needed later, the optimization parameter is used for prediction.
  • the stochastic gradient descent method can train the optimal parameters in a small batch manner.
  • the batch size is 128.
  • the target background application is currently in the background and is not currently called.
  • Forming a binary vector then inputting a reference module, combining the reference model with the optimized parameters obtained before training, performing prediction according to the binary vector, and then generating a prediction result, and finally controlling the target background application according to the prediction result, such as closing or maintaining the Target background application.
  • the training process of the reference model can be completed on the server side or on the electronic device side.
  • the training process and the actual prediction process of the reference model are completed on the server side, when the optimized reference model needs to be used, the usage status of the preset background application before the current time can be input to the server, and the server actually After the prediction is completed, the prediction result is sent to the electronic device side, and the electronic device controls the preset background application according to the prediction result.
  • the usage state of the plurality of time periods before the current time of the preset background application may be input to the electronic device.
  • the electronic device controls the preset background application according to the predicted result.
  • the preset background application can be used for multiple time periods before the current time.
  • the status is input to the electronic device, and after the actual prediction of the electronic device is completed, the electronic device controls the preset background application according to the predicted result.
  • the trained reference model file model file
  • the trained reference model file model file
  • the usage status of the preset background application before the current time is obtained. Enter the trained reference model file (model file) and calculate to get the predicted value.
  • the background application control method of the embodiment of the present application is applied to an electronic device, and obtains a plurality of continuous use states of a preset background application in a preset time period, and forms a plurality of consecutive use states.
  • a binary vector a binary vector
  • the first binary vector is trained as a training data input reference model to obtain an optimized parameter of the trained reference model
  • a plurality of continuous use states corresponding to the current time of the preset background application are obtained, and Forming a second binary vector corresponding to a plurality of consecutive use states of the current time, finally predicting according to the reference model, the optimized parameter, and the second binary vector, generating a prediction result, and controlling the preset background application according to the prediction result .
  • It can improve the accuracy of predicting whether the preset background application entering the background still needs to be used, thereby improving the intelligence and accuracy of controlling the incoming application into the background, and reducing the memory occupation.
  • FIG. 4 is another schematic flowchart of an application management and control method according to an embodiment of the present application.
  • the specific process of the application management and control method in the embodiment of the present application is described by using the background application control device integrated in the electronic device as an example.
  • 203, 204, and 205 in FIG. 4 are the same as 101, 102, and 103 in FIG. 1, respectively, and are not described herein again.
  • the preset statistical period may be one or more hours, one or more days, and in order to obtain more accurate prediction results, the more training data, the better.
  • the plurality of background applications may be scored according to the running data of the background application based on the third preset formula, wherein the third preset formula is:
  • memory i is the memory usage ratio
  • couter i is the total usage time
  • t i is the total usage time
  • seq i is the latest usage order of the default background application
  • Group is the grouping coefficient
  • G f is the sinking factor
  • T f is Time weakening factor.
  • the value of memory i is between 0-1.
  • Seq i is the latest usage order of the default background application. For example, the order value of the currently used application is equal to 0, the application order value used last is 1, and the previous application is 2, and so on.
  • Group is a grouping coefficient
  • Group is greater than or equal to 1, less than or equal to 10, because each background application is different, the user is different, and the selected grouping coefficients are also different.
  • the prediction result of different grouping coefficients of each background application may be calculated in advance, and when the prediction result of a certain grouping coefficient is more accurate, the grouping coefficient is adopted.
  • G f is the sinking factor, G f is greater than 1, less than or equal to 3, after using the sinking factor, the denominator is increased, thereby reducing the score value, and reducing the influence of seq i on the score.
  • T f is a time weakening factor, T f is greater than or equal to 0, less than or equal to 1, after using the time weakening factor, the molecule is reduced, thereby reducing the score value, and reducing the influence of t i on the score. Preventing the weight of a parameter from being too large, has an excessive impact on the score, and makes the weight between the data more reasonable.
  • a preset score threshold may be set in advance, such as according to a percentage of the highest score of the score, such as a preset score threshold of 80% of the highest score of the score. After the electronic device is used, according to the current electronic device usage, a new preset score threshold is set to replace the previous preset score threshold, and an average or typical value may also be counted through the network.
  • the preset scoring threshold can be reacquired after a period of use or by re-acquiring instructions.
  • the background application control method of the embodiment of the present application is applied to an electronic device, and first selects a part of the background application as a preset background application, and then obtains a plurality of consecutive use states of the preset background application within a preset time period.
  • the embodiment of the present application further provides a background application control device.
  • the meaning of the noun is the same as the application management method described above. For specific implementation details, refer to the description in the method embodiment.
  • FIG. 5 is a schematic structural diagram of a background application control device according to an embodiment of the present application.
  • the background application management device 300 is applied to an electronic device, and the background application management device 300 includes an obtaining unit 301, a training unit 302, and a control unit 303, where:
  • the obtaining unit 301 is configured to acquire a plurality of consecutive use states of the preset background application in a preset time period, and form the acquired multiple continuous use states into a first binary value vector.
  • the preset time period can be set according to the usage period, such as 8:00 to 12:00 in the morning, 6 to 10 in the evening, or Monday to Friday, and Saturday to Sunday.
  • the preset time period can be several hours or days, or even longer.
  • a plurality of continuous use states of the preset background application are obtained in the preset time period, and the use state is whether the preset background application is in use, if the use is recorded as 1, and if not used, the record is 0.
  • the usage status is continuous. If the usage status is recorded every 5 minutes within one hour, 12-13 consecutive use statuses can be obtained.
  • the judgment criterion of the usage status can be as long as the user can call once within 5 minutes. I think it is in use. Of course, the interval can be set to other values, such as 10 minutes, 3 minutes, and so on.
  • the default background application can be an application for web browsing, an application for video playback, an application for news browsing, and the like.
  • the data of the first binary vector is a set of binary data, such as ⁇ 0, 1, 1, 0, 0, 0, 1 ⁇ , wherein the data 0 indicates that the preset background application is not called during the time period, and the data 1 Indicates that the default background application is called at least once during this time period.
  • the training unit 302 is configured to acquire a reference model, and the first binary vector is used as the training data input reference model for training, and the optimized parameters of the trained reference model are obtained.
  • the reference model can be a cyclic neural network model.
  • the reference model can be a hybrid neural network model, a Gaussian mixture model or a convolutional neural network model.
  • unit as used herein may be taken to mean a software object that is executed on the computing system.
  • the units described herein can be considered as implementation objects on the computing system.
  • the apparatus and method described herein may be implemented in software, and may of course be implemented in hardware, all of which are within the scope of the present application.
  • FIG. 6 is another schematic structural diagram of a background application control device according to an embodiment of the present application.
  • the cyclic neural network model includes a network sub-unit 3021, a classifier 3022, a loss value calculator 3023, and an optimization sub-unit 3024. among them:
  • the network subunit 3021 is configured to input the first binary vector as the training data into the network subunit to obtain an intermediate value.
  • the network sub-unit 3021 specifically includes an input layer, a hidden layer, and an output layer.
  • the first binary vector is used as the input layer of the training data input network subunit, the input layer inputs the first binary vector into the hidden layer, and the hidden layer calculates the intermediate value by cyclically calculating the first binary vector, and outputs the intermediate value through the output layer. .
  • the classifier 3022 is configured to input the intermediate value into the classifier to obtain a probability corresponding to the plurality of prediction results.
  • the classifier 3022 can input the intermediate value into the classifier based on the first preset formula to obtain a probability corresponding to the plurality of prediction results, where the first preset formula is:
  • Z K is the target intermediate value
  • C is the number of categories of the prediction result
  • the true value may be 0 or 1
  • Z j is the jth intermediate value.
  • the median value of the target is the median value of the kth.
  • the prediction result is set according to requirements. For example, if the background application's prediction result includes cleaning and not cleaning, the number of categories C of the prediction result is 2.
  • the loss value calculator 3023 is configured to obtain a loss value according to the plurality of prediction results and a plurality of probabilities corresponding thereto.
  • the loss value calculator 3023 may obtain a loss value according to the plurality of prediction results and a plurality of probabilities corresponding thereto according to the fourth preset formula, wherein the fourth preset formula is:
  • the loss value calculator 3023 may obtain a loss value according to the plurality of prediction results and a plurality of probabilities corresponding thereto based on the second preset formula, wherein the second preset formula is:
  • C is the number of categories of prediction results
  • y k is the true value
  • E is the average value
  • the loss value obtained by the fourth preset formula can be obtained, and then the average value is obtained by averaging, and the data is more accurate.
  • the acquisition loss value may be acquired according to the cross entropy or the like in addition to the above method.
  • the optimization subunit 3024 is configured to perform training according to the loss value to obtain an optimization parameter.
  • the optimization sub-unit 3024 can be trained using a stochastic gradient descent method based on the loss value. Training can also be performed according to the batch gradient descent method or the gradient descent method.
  • Training is performed using the stochastic gradient descent method, and the training can be completed when the loss value is equal to or less than the preset loss value. It is also possible to complete the training when there are no changes in the two or more loss values continuously acquired. Of course, it is also possible to directly set the number of iterations of the random gradient descent method according to the loss value. After the number of iterations is completed, the training is completed. After the training is completed, each parameter of the reference model at this time is obtained, and the each parameter is saved as an optimization parameter, and when the prediction is needed later, the optimization parameter is used for prediction.
  • the stochastic gradient descent method can train the optimal parameters in a small batch manner.
  • the batch size is 128.
  • the control unit 303 is configured to acquire a plurality of consecutive use states corresponding to the current time of the preset background application, and form a plurality of consecutive use states corresponding to the current time to form a second binary vector according to the reference model, the optimization parameter, and the second The binary vector is predicted, the predicted result is generated, and the preset background application is controlled according to the predicted result.
  • the target background application If it is necessary to determine whether the target background application can be cleaned up, first obtain a plurality of consecutive usage states of the preset background application corresponding to the current time. Because it is a background application, the target background application is currently in the background and is not currently called. Obtain the usage status of the first few time periods of the current time. For example, the current time is used as the reference, and every 5 minutes is used as a time period to obtain the usage status of the first five consecutive time segments, and then the five usage states are formed.
  • a binary vector then input the reference module, the reference model combines the optimized parameters obtained before training, predicts according to the binary vector, and then generates a prediction result, and finally controls the target background application according to the prediction result, such as closing or maintaining the target Background application.
  • the training process of the reference model can be completed on the server side or on the electronic device side.
  • the training process and the actual prediction process of the reference model are completed on the server side, when the optimized reference model needs to be used, the usage status of the preset background application before the current time can be input to the server, and the server actually After the prediction is completed, the prediction result is sent to the electronic device side, and the electronic device controls the preset background application according to the prediction result.
  • the usage state of the plurality of time periods before the current time of the preset background application may be input to the electronic device.
  • the electronic device controls the preset background application according to the predicted result.
  • FIG. 7 is a schematic diagram of another application scenario of a background application control device according to an embodiment of the present disclosure.
  • the training process of the reference model is completed on the server side
  • the actual prediction process of the reference model is completed on the electronic device side
  • the optimized reference model needs to be used
  • the preset background application can be used for multiple time periods before the current time.
  • the status is input to the electronic device, and after the actual prediction of the electronic device is completed, the electronic device controls the preset background application according to the predicted result.
  • the trained reference model file model file
  • the smart device If it is necessary to determine whether the current background application can be cleaned, the usage status of the background application before the current time is preset, and the input state is input. Go to the trained reference model file (model file) and calculate to get the predicted value.
  • the background application control device of the embodiment of the present application is applied to an electronic device, and obtains a plurality of continuous use states of a preset background application in a preset time period, and forms a plurality of continuous use states.
  • a binary vector a binary vector
  • the first binary vector is trained as a training data input reference model to obtain an optimized parameter of the trained reference model
  • a plurality of continuous use states corresponding to the current time of the preset background application are obtained, and Forming a second binary vector corresponding to a plurality of consecutive use states of the current time, finally predicting according to the reference model, the optimized parameter, and the second binary vector, generating a prediction result, and controlling the preset background application according to the prediction result .
  • It can improve the accuracy of predicting whether the preset background application entering the background still needs to be used, thereby improving the intelligence and accuracy of controlling the incoming application into the background, and reducing the memory occupation.
  • FIG. 8 is still another schematic structural diagram of a background application control device according to an embodiment of the present application.
  • the background application management device of the embodiment of the present application includes: a scoring unit 304, a setting unit 305, an obtaining unit 301, a training unit 302, and a management unit 303. among them:
  • the scoring unit 304 is configured to score a plurality of background applications according to the running data of the background application in a preset statistical period.
  • the preset statistical period can be one or more hours, one or more days, and in order to obtain more accurate prediction results, the more training data, the better.
  • the scoring unit 304 may score a plurality of background applications according to the running data of the background application based on the third preset formula, wherein the third preset formula is:
  • memory i is the memory usage ratio
  • couter i is the total usage time
  • t i is the total usage time
  • seq i is the latest usage order of the default background application
  • Group is the grouping coefficient
  • G f is the sinking factor
  • T f is Time weakening factor.
  • the value of memory i is between 0-1, and seq i is the latest usage order of the default background application. For example, the order value of the currently used application is equal to 0, and the last used application order value is 1, and then the previous one is used. The application is 2, and so on.
  • Group is a grouping coefficient
  • Group is greater than or equal to 1, less than or equal to 10, because each background application is different, the user is different, and the selected grouping coefficients are also different.
  • the prediction result of different grouping coefficients of each background application may be calculated in advance, and when the prediction result of a certain grouping coefficient is more accurate, the grouping coefficient is adopted.
  • G f is the sinking factor, G f is greater than 1, less than or equal to 3, after using the sinking factor, the denominator is increased, thereby reducing the score value, and reducing the influence of seq i on the score.
  • T f is a time weakening factor, T f is greater than or equal to 0, less than or equal to 1, after using the time weakening factor, the molecule is reduced, thereby reducing the score value, and reducing the influence of t i on the score. Preventing the weight of a parameter from being too large, has an excessive impact on the score, and makes the weight between the data more reasonable.
  • the setting unit 305 is configured to set a background application whose score exceeds a preset rating threshold as a preset background application.
  • a preset rating threshold can be set in advance, such as based on the percentage of the highest score of the rating, such as the preset rating threshold being 80% of the highest score of the rating. After the electronic device is used, according to the current electronic device usage, a new preset score threshold is set to replace the previous preset score threshold, and an average or typical value may also be counted through the network.
  • the preset scoring threshold can be reacquired after a period of use or by re-acquiring instructions.
  • the background application control device of the embodiment of the present application is applied to an electronic device, and first selects a part of the background application as a preset background application, and then obtains a plurality of consecutive use states of the preset background application within a preset time period.
  • the background application control device of the embodiment of the present application is applied to an electronic device, and obtains a plurality of continuous use states of a preset background application in a preset time period, and forms a plurality of acquired continuous use states.
  • a first binary vector is trained as a training data input reference model to obtain an optimized parameter of the trained reference model; and the plurality of consecutive use states corresponding to the current time is formed into a second binary vector, according to
  • the reference model, the optimization parameters, and the second binary vector are used for prediction, the prediction result is generated, and the preset background application is controlled according to the prediction result. It can improve the accuracy of forecasting the default background application, thus improving the accuracy of controlling the application entering the background.
  • the background application control device and the background application control method in the above embodiment are in the same concept, and any method provided in the embodiment of the background application management and control method may be run on the background application control device.
  • any method provided in the embodiment of the background application management and control method may be run on the background application control device.
  • details of the implementation process refer to the embodiment of the background application management and control method, which is not described here.
  • the electronic device 400 includes a processor 401 and a memory 402.
  • the processor 401 is electrically connected to the memory 402.
  • the processor 400 is a control center of the electronic device 400 that connects various portions of the entire electronic device using various interfaces and lines, executes the electronic by running or loading a computer program stored in the memory 402, and recalling data stored in the memory 402.
  • the memory 402 can be used to store software programs and modules, and the processor 401 executes various functional applications and data processing by running computer programs and modules stored in the memory 402.
  • the memory 402 can mainly include a storage program area and a storage data area, wherein the storage program area can store an operating system, a computer program required for at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area can be stored according to Data created by the use of electronic devices, etc.
  • memory 402 can include high speed random access memory, and can also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device. Accordingly, memory 402 can also include a memory controller to provide processor 401 access to memory 402.
  • the processor 401 in the electronic device 400 loads the instructions corresponding to the process of one or more computer programs into the memory 402 according to the following flow, and is stored and stored in the memory 402 by the processor 401.
  • the computer program in which to implement various functions, as follows:
  • the reference model is trained to obtain optimization parameters of the reference model after training; obtaining a plurality of consecutive use states corresponding to the current time by the preset background application, and forming a second binary vector by using a plurality of consecutive use states corresponding to the current time.
  • the prediction is performed according to the reference model, the optimization parameter, and the second binary vector, the prediction result is generated, and the preset background application is controlled according to the prediction result.
  • the processor 401 is further configured to:
  • the cyclic neural network model is selected as the reference model.
  • the processor 401 is further configured to:
  • the cyclic neural network model includes a network subunit and a classifier
  • the first binary vector is used as the training data input reference model for training, and the optimized parameters of the reference model after training include:
  • Training is performed according to the loss value to obtain optimized parameters.
  • the processor 401 is further configured to:
  • Z K is the target intermediate value
  • C is the number of categories of prediction results
  • Z j is the jth intermediate value
  • the processor 401 is further configured to:
  • the processor 401 is further configured to:
  • E is the average.
  • the processor 401 is further configured to:
  • Training is performed using a stochastic gradient descent method based on the loss value.
  • the processor 401 is further configured to:
  • the processor 401 is further configured to:
  • the plurality of background applications are scored according to the running data of the background application based on the third preset formula, wherein the third preset formula is:
  • memory i is the memory usage ratio
  • couter i is the total usage time
  • t i is the total usage time
  • seq i is the latest usage order of the default background application
  • Group is the grouping coefficient
  • Gf is the sinking factor
  • Tf is the time. Weakening factor.
  • the electronic device obtains a plurality of consecutive use states of the preset background application in a preset time period, and forms the obtained multiple continuous use states into a first binary value vector;
  • the first binary vector is trained as a training data input reference model to obtain an optimized parameter of the trained reference model;
  • a plurality of consecutive use states corresponding to the current time are formed into a second binary vector, according to the reference model, the optimized parameter, and the first
  • the binary value vector is predicted, the prediction result is generated, and the preset background application is controlled according to the prediction result. It can improve the accuracy of forecasting the default background application, thus improving the accuracy of controlling the application entering the background.
  • the electronic device 400 may further include: a display 403, a radio frequency circuit 404, an audio circuit 405, and a power source 406.
  • the display 403, the radio frequency circuit 404, the audio circuit 405, and the power source 406 are electrically connected to the processor 401, respectively.
  • Display 403 can be used to display information entered by a user or information provided to a user, as well as various graphical user interfaces, which can be composed of graphics, text, icons, video, and any combination thereof.
  • the display 403 can include a display panel.
  • the display panel can be configured in the form of a liquid crystal display (LCD), or an organic light-emitting diode (OLED).
  • LCD liquid crystal display
  • OLED organic light-emitting diode
  • the radio frequency circuit 404 can be used to transmit and receive radio frequency signals to establish wireless communication with network devices or other electronic devices through wireless communication, and to transmit and receive signals with network devices or other electronic devices.
  • the audio circuit 405 can be used to provide an audio interface between the user and the electronic device through the speaker and the microphone.
  • Power source 406 can be used to power various components of electronic device 400.
  • the power supply 406 can be logically coupled to the processor 401 through a power management system to enable functions such as managing charging, discharging, and power management through the power management system.
  • the electronic device 400 may further include a camera, a Bluetooth module, and the like, and details are not described herein again.
  • the embodiment of the present application further provides a storage medium, where the storage medium stores a computer program, and when the computer program runs on the computer, causes the computer to execute the application management and control method in any of the above embodiments, for example, acquiring a preset background application.
  • the storage medium may be a magnetic disk, an optical disk, a read only memory (ROM), or a random access memory (RAM).
  • ROM read only memory
  • RAM random access memory
  • a general tester in the field can understand all or part of the process of implementing the background application control method of the embodiment of the present application, and the related program can be controlled by a computer program.
  • the computer program can be stored in a computer readable storage medium, such as in a memory of the electronic device, and executed by at least one processor in the electronic device, and can include, for example, a background application during execution.
  • the storage medium may be a magnetic disk, an optical disk, a read only memory, a random access memory, or the like.
  • each functional module may be integrated into one processing chip, or each module may exist physically separately, or two or more modules may be integrated into one module.
  • the above integrated modules can be implemented in the form of hardware or in the form of software functional modules.
  • An integrated module, if implemented in the form of a software functional module and sold or used as a standalone product, may also be stored in a computer readable storage medium such as a read only memory, a magnetic disk or an optical disk.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Quality & Reliability (AREA)
  • Software Systems (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本申请公开了一种后台应用程序管控方法、存储介质及电子设备，获取预设后台应用程序的使用状态形成的第一二值向量；将第一二值向量输入参考模型进行训练得到优化参数；获取预设后台应用程序当前时间的使用状态形成的第二二值向量，根据参考模型、优化参数以及第二二值向量进行预测，进而对预设后台应用程序进行管控。

Description

后台应用程序管控方法、存储介质及电子设备
本申请要求于2017年09月30日提交中国专利局、申请号为201710922671.7、申请名称为“后台应用程序管控方法、装置、存储介质及电子设备”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请属于通信技术领域，尤其涉及一种后台应用程序管控方法、存储介质及电子设备。
背景技术
随着电子技术的发展，人们通常在电子设备上安装很多应用程序。当用户在电子设备中打开多个应用程序时，多个应用程序会在电子设备的后台运行。这些后台运行的应用程序会占用电子设备的内存，并且增加功耗。
发明内容
本申请提供一种后台应用程序管控方法、存储介质及电子设备，能够提升对应用程序进行管控的准确性。
第一方面，本申请实施例提供一种后台应用程序管控方法，应用于电子设备，该方法包括：
获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续所述使用状态形成第一二值向量；
获取参考模型，并将所述第一二值向量作为训练数据输入所述参考模型进行训练，得到训练后的所述参考模型的优化参数；
获取所述预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续所述使用状态形成第二二值向量，根据所述参考模型、所述优化参数以及所述第二二值向量进行预测，生成预测结果，并根据所述预测结果对所述预设后台应用程序进行管控。
第二方面，本申请实施例提供一种存储介质，其上存储有计算机程序，当所述计算机程序在计算机上运行时，使得所述计算机执行上述的后台应用程序管控方法。
第三方面，本申请实施例提供一种电子设备，包括处理器和存储器，所述存储器有计算机程序，所述处理器通过调用所述计算机程序，其中，所述处理器还执行：
获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续所述使用状态形成第一二值向量；
获取参考模型，并将所述第一二值向量作为训练数据输入所述参考模型进行训练，得到训练后的所述参考模型的优化参数；
获取所述预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续所述使用状态形成第二二值向量，根据所述参考模型、所述优化参数以及所述第二二值向量进行预测，生成预测结果，并根据所述预测结果对所述预设后台应用程序进 行管控。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍。显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的后台应用程序管控装置的系统示意图；
图2为本申请实施例提供的后台应用程序管控装置的应用场景示意图；
图3为本申请实施例提供的后台应用程序管控方法的流程示意图；
图4为本申请实施例提供的后台应用程序管控方法的另一流程示意图；
图5为本申请实施例提供的后台应用程序管控装置的结构示意图；
图6为本申请实施例提供的后台应用程序管控装置的另一结构示意图；
图7为本申请实施例提供的后台应用程序管控装置的另一应用场景示意图；
图8为本申请实施例提供的后台应用程序管控装置的又一结构示意图；
图9为本申请实施例提供的电子设备的结构示意图；
图10为本申请实施例提供的电子设备的另一结构示意图。
具体实施方式
请参照图式，其中相同的组件符号代表相同的组件，本申请的原理是以实施在一适当的运算环境中来举例说明。以下的说明是基于所例示的本申请具体实施例，其不应被视为限制本申请未在此详述的其它具体实施例。
本申请实施例提供一种后台应用程序管控方法，应用于电子设备，所述方法包括：
获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续所述使用状态形成第一二值向量；
获取参考模型，并将所述第一二值向量作为训练数据输入所述参考模型进行训练，得到训练后的所述参考模型的优化参数；
获取所述预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续所述使用状态形成第二二值向量，根据所述参考模型、所述优化参数以及所述第二二值向量进行预测，生成预测结果，并根据所述预测结果对所述预设后台应用程序进行管控。
在所述的后台应用程序管控方法中，所述参考模型为循环神经网络模型。
在所述的后台应用程序管控方法中，所述循环神经网络模型包括网络子单元和分类器；
所述将所述第一二值向量作为训练数据输入所述参考模型进行训练，得到训练后的所述参考模型的优化参数具体包括：
将第一二值向量作为训练数据输入所述网络子单元得到中间值；
将所述中间值输入所述分类器得到对应多个所述预测结果的概率；
根据多个所述预测结果和与其对应的多个所述概率得到损失值；
根据所述损失值进行训练，得到所述优化参数。
在所述的后台应用程序管控方法中，所述将所述中间值输入所述分类器得到对应多个 所述预测结果的概率具体包括：
基于第一预设公式将所述中间值输入所述分类器得到对应多个所述预测结果的概率，其中所述第一预设公式为：
Figure PCTCN2018102187-appb-000001
其中，Z K为目标中间值，C为预测结果的类别数，Z j为第j个中间值。
在所述的后台应用程序管控方法中，所述根据多个所述预测结果和与其对应的多个所述概率得到损失值具体包括：
基于第二预设公式根据多个所述预测结果和与其对应的多个所述概率得到损失值，其中所述第二预设公式为：
Figure PCTCN2018102187-appb-000002
其中y k为真实值，E为求平均值。
在所述的后台应用程序管控方法中，所述根据多个所述预测结果和与其对应的多个所述概率得到损失值具体包括：
基于第四预设公式根据多个所述预测结果和与其对应的多个所述概率得到损失值，其中所述第四预设公式为：
Figure PCTCN2018102187-appb-000003
其中
Figure PCTCN2018102187-appb-000004
为第一预设公式的结果，E为求平均值。
在所述的后台应用程序管控方法中，所述根据多个所述预测结果和与其对应的多个所述概率得到损失值具体包括：
基于交叉熵模型，根据多个所述预测结果和与其对应的多个所述概率得到损失值。
在所述的后台应用程序管控方法中，所述根据所述损失值进行训练具体包括：
根据所述损失值利用随机梯度下降法进行训练。
在所述的后台应用程序管控方法中，所述获取预设后台应用程序在预设时间段内的多个连续使用状态之前，所述方法还包括：
在预设统计周期内，根据后台应用程序的运行数据对多个后台应用程序进行评分；
将评分超过预设评分阈值的后台应用程序设为所述预设后台应用程序。
在所述的后台应用程序管控方法中，所述根据后台应用程序的运行数据对多个后台应用程序进行评分具体包括：
基于第三预设公式根据后台应用程序的运行数据对多个后台应用程序进行评分，其中所述第三预设公式为：
Figure PCTCN2018102187-appb-000005
其中memory i为内存占用比，couter i为总使用次数，t i为总使用时间，seq i为预设后台应用程序的最近使用次序，Group为分组系数，G f为下沉因子，T f为时间弱化因子。
在所述的后台应用程序管控方法中，根据所述预测结果对所述预设后台应用程序进行管控具体包括：
当预测结果为关闭时，将所述预设后台应用程序关闭；
当预测结果为保留时，将所述预设后台应用程序保留。
在本文中提及“实施例”意味着，结合实施例描述的特定特征、结构或特性可以包含在本申请的至少一个实施例中。在说明书中的各个位置出现该短语并不一定均是指相同的实施例，也不是与其它实施例互斥的独立的或备选的实施例。本领域技术人员显式地和隐式地理解的是，本文所描述的实施例可以与其它实施例相结合。
请参阅图1，图1为本申请实施例提供的后台应用程序管控装置的系统示意图。该后台应用程序管控装置主要用于：获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续使用状态形成第一二值向量；获取参考模型，并将第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续使用状态形成第二二值向量，根据参考模型、优化参数以及第二二值向量进行预测，生成预测结果，然后根据该预测结果判断该预设后台应用程序是否需要被使用，以对预设后台应用程序进行管控，例如关闭、或者冻结等。
具体的，请参阅图2，图2为本申请实施例提供的后台应用程序管控装置的应用场景示意图。比如，后台应用程序管控装置在接收到管控请求时，检测到在电子设备的后台运行的应用程序包括预设后台应用程序a、预设后台应用程序b以及预设后台应用程序c；然后分别获取预设后台应用程序a对应的参考模型A、预设后台应用程序b对应的参考模型B以及预设后台应用程序c对应的参考模型C；通过参考模型A对预设后台应用程序a是否需要被使用的概率进行预测，得到概率a’，通过参考模型B对预设后台应用程序b是否需要被使用的概率进行预测，得到概率b’，参考模型C对预设后台应用程序c是否需要被使用的概率进行预测，得到概率c’；根据概率a’、概率b’以及概率c’对后台运行的预设后台应用程序a、预设后台应用程序b以及预设后台应用程序c进行管控，例如将概率最低的预设后台应用程序b关闭。
本申请实施例提供一种后台应用程序管控方法，该后台应用程序管控方法的执行主体可以是本申请实施例提供的后台应用程序管控装置，或者集成了该后台应用程序管控装置的电子设备，其中该后台应用程序管控装置可以采用硬件或者软件的方式实现。
本申请实施例将从后台应用程序管控装置的角度进行描述，该后台应用程序管控装置具体可以集成在电子设备中。该后台应用程序管控方法包括：获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续使用状态形成第一二值向量；获取参考模型，并将第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前 时时间的多个连续使用状态形成第二二值向量，根据参考模型、优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。
请参阅图3，图3为本申请实施例提供的后台应用程序管控方法的流程示意图。本申请实施例提供的后台应用程序管控方法应用于电子设备，具体流程可以如下：
101，获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续使用状态形成第一二值向量。
具体的，预设时间段可以根据使用周期来设置，如早上8点到12点，晚上6点到10点，也可以为周一至周五，周六至周日。预设时间段可以为几个小时或几天，甚至更长的时间。然后在该预设时间段内获取预设后台应用程序的多个连续使用状态，使用状态为该预设后台应用程序是否在使用，如果在使用则记录为1，没有使用则记录为0。并且使用状态为连续的，如一个小时内，每隔5分钟记录一次使用状态，则可以获取12-13个连续使用状态，使用状态的判断标准可以为，5分钟内只要被用户调用一次即可认为在使用。当然间隔时间可以设置为其他值，如10分钟、3分钟等。预设后台应用程序可以为网页浏览的应用程序、视频播放的应用程序、新闻浏览的应用程序等。
第一二值向量的数据为二进制数据的集合，如{0,1，1,0,0,0,1}，其中的数据0标示该时间段内预设后台应用程序未被调用，数据1标示该时间段内预设后台应用程序至少被调用1次。
102，获取参考模型，并将第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数。
其中，参考模型可以为循环神经网络模型。当然参考模型可以为混合神经网络模型、高斯混合模型或卷积神经网络模型等。
循环神经网络模型包括网络子单元和分类器。具体的，将第一二值向量作为训练数据输入网络子单元得到中间值；将中间值输入分类器得到对应多个预测结果的概率；根据多个预测结果和与其对应的多个概率得到损失值；根据损失值进行训练，得到优化参数。
其中网络子单元具体包括输入层、隐层和输出层。将第一二值向量作为训练数据输入网络子单元的输入层，输入层将第一二值向量输入隐层，隐层将第一二值向量循环计算得到中间值，再将中间值通过输出层输出。
在一些实施方式中，可以基于第一预设公式将中间值输入分类器得到对应多个预测结果的概率，其中第一预设公式为：
Figure PCTCN2018102187-appb-000006
其中，Z K为目标中间值，C为预测结果的类别数，Z j为第j个中间值。目标中间值为第k个的中间值。预测结果可以根据需要设置，如后台应用程序的预测结果包括清理和不清理两类，则预测结果的类别数C为2。
在一些实施方式中，损失值可以基于第四预设公式根据多个预测结果和与其对应的多个概率得到损失值，其中第四预设公式为：
Figure PCTCN2018102187-appb-000007
其中C为预测结果的类别数，y k为真实值，真实值可以为0或1，
Figure PCTCN2018102187-appb-000008
为第一预设公式的结果。
在一些实施方式中，可以基于第二预设公式根据多个预测结果和与其对应的多个概率得到损失值，其中第二预设公式为：
Figure PCTCN2018102187-appb-000009
其中C为预测结果的类别数，y k为真实值，真实值可以为0或1，
Figure PCTCN2018102187-appb-000010
为第一预设公式的结果。E为求平均值，可以获取多个第四预设公式获取的损失值，然后求平均值获取平均后的损失值，数据更准确。
本申请中的术语“第一”、“第二”等是用于区别不同对象，而不是用于描述特定顺序。
需要说明的是，获取损失值除了上述方法外，还可以根据交叉熵等方式获取。
在一些实施方式中，可以根据损失值利用随机梯度下降法进行训练。还可以根据批量梯度下降法或梯度下降法进行训练。
利用随机梯度下降法进行训练，可以当损失值等于或小于预设损失值时，则完成训练。也可以当连续获取的两个或多个损失值没有变化时，则完成训练。当然还可以不根据损失值，直接设定随机梯度下降法的迭代次数，迭代次数完成后，则完成训练。训练完成后，获取此时的参考模型的各个参数，并将该各个参数保存为优化参数，后续需要预测时，使用该优化参数进行预测。
其中随机梯度下降法可以采用小批量的方式训练得到最优参数。如批量大小为128。
103，获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续使用状态形成第二二值向量，根据参考模型、优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。
若需要判断预设后台应用程序是否可清理，首先获取该预设后台应用程序对应当前时间的多个连续使用状态，因为是后台应用程序，所以目标后台应用程序当前在后台中，当前没有被调用，获取当前时间的前几个时间段的使用状态，如以当前时刻为基准，往前每隔5分钟为一个时间段，获取前面5个连续时间段的使用状态，然后将这5个使用状态形成一个二值向量，然后输入参考模块，参考模型结合之前训练得到的优化参数，根据这个二值向量进行预测，然后生成预测结果，最后根据预测结果管控该目标后台应用程序，如关闭或保持该目标后台应用程序。
需要说明的是，参考模型的训练过程可以在服务器端也可以在电子设备端完成。当参考模型的训练过程、实际预测过程都在服务器端完成时，需要使用优化后的参考模型时，可以将预设后台应用程序当前时间前的多个时间段的使用状态输入到服务器，服务器实际预测完成后，将预测结果发送至电子设备端，电子设备再根据预测结果管控该预设后台应用程序。
当参考模型的训练过程、实际预测过程都在电子设备端完成时，需要使用优化后的参 考模型时，可以将预设后台应用程序当前时间前的多个时间段的使用状态输入到电子设备，电子设备实际预测完成后，电子设备根据预测结果管控该预设后台应用程序。
当参考模型的训练过程在服务器端完成，参考模型的实际预测过程在电子设备端完成时，需要使用优化后的参考模型时，可以将预设后台应用程序当前时间前的多个时间段的使用状态输入到电子设备，电子设备实际预测完成后，电子设备根据预测结果管控该预设后台应用程序。可选的，可以将训练好的参考模型文件(model文件)移植到智能设备上，若需要判断当前后台应用是否可清理，则获取预设后台应用程序当前时间前的多个时间段的使用状态，输入到训练好的参考模型文件(model文件)，计算即可得到预测值。
上述所有的技术方案，可以采用任意结合形成本申请的可选实施例，在此不再一一赘述。
由上述可知，本申请实施例的后台应用程序管控方法，应用于电子设备，通过获取预设后台应用程序在预设时间段内的多个连续使用状态，并将多个连续使用状态形成第一二值向量；再将该第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；接着获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续使用状态形成第二二值向量，最后根据参考模型、所优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。可以提高对进入后台的预设后台应用程序是否还需要使用的概率进行预测的准确性，从而提升对进入后台的预测应用程序进行管控的智能化和准确性，减少内存占用。
请参阅图4，图4为本申请实施例提供的应用程序管控方法的另一流程示意图。本申请实施例以后台应用程序管控装置集成在电子设备为例，描述本申请实施例的应用程序管控方法的具体流程。其中，图4中的203、204及205分别与图1中的101、102及103相同，在此不再赘述。
201，在预设统计周期内，根据后台应用程序的运行数据对多个后台应用程序进行评分。
其中，预设统计周期可以为一个或多个小时，一天或多天，为了获取更准确的预测结果，训练数据越多越好。
可以基于第三预设公式根据后台应用程序的运行数据对多个后台应用程序进行评分，其中第三预设公式为：
Figure PCTCN2018102187-appb-000011
其中memory i为内存占用比，couter i为总使用次数，t i为总使用时间，seq i为预设后台应用程序的最近使用次序，Group为分组系数，G f为下沉因子，T f为时间弱化因子。memory i的取值在0-1之间。seq i为预设后台应用程序的最近使用次序，例如当前使用的应用的次序值等于0，上一个使用的应用次序值为1，再之前使用的一个应用为2，依次类推。Group为分组系数，Group大于等于1，小于等于10，因为每个后台应用程序不同，用户也不同，选取的分组系数也不同。例如，可以预先计算每个后台应用程序不同分组系数的预测结果，当某个分组系数的预测结果更准确时，则采用该分组系数。G f为下沉因子，G f大于1，小于等于3，使用下沉因子后，增大了分母，进而减小了评分值，减小了seq i对评分的影响。 T f为时间弱化因子，T f大于等于0，小于等于1，使用时间弱化因子后，减小了分子，进而减小了评分值，减小了t i对评分的影响。防止一个参数的权重太大，对评分有过重的影响，使各个数据之间的权重更合理。
202，将评分超过预设评分阈值的后台应用程序设为预设后台应用程序。
可以预先设置一个预设评分阈值，如根据评分的最高分的百分比来设置，如预设评分阈值为评分的最高分的80％。也可以电子设备使用后，根据当前电子设备的使用情况，设定新的预设评分阈值替换之前的预设评分阈值，也可以通过网络统计一个平均值或典型值。
也可以将多个后台应用程序按照评分从高到低排列形成列表，将列表前几名的后台应用程序，如前5名，设置为预设后台应用程序。也可以将列表的某个排名的后台应用程序的评分设为预设评分阈值，如排名第6的后台应用程序的评分设为预设评分阈值。也可以不同条件下获取多个列表，然后取平均值。如每次都是取第6名的后台应用程序的评分，然后求平均值得到预设评分阈值。当然，使用一段时间后或经重新获取指令可以重新获取预设评分阈值。
考虑到时间序列分析是针对单个应用程序进行的，分析电子设备所有的应用程序数据量太大，实用性不强，因此我们要根据使用习惯选取排名较高的多个应用程序，并且清理这几个应用程序确实能大大释放内存。
本申请实施例的后台应用程序管控方法，应用于电子设备，首先选取部分后台应用程序为预设后台应用程序，然后通过获取预设后台应用程序在预设时间段内的多个连续使用状态，并将多个连续使用状态形成第一二值向量；再将该第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；接着获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续使用状态形成第二二值向量，最后根据参考模型、所优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。可以提高对进入后台的预设后台应用程序是否还需要使用的概率进行预测的准确性，从而提升对进入后台的预测应用程序进行管控的智能化和准确性，减少内存占用。
为便于更好地实施本申请实施例提供的应用程序管控方法，本申请实施例还提供一种后台应用程序管控装置。其中名词的含义与上述应用程序管控方法相同，具体实现细节可以参考方法实施例中的说明。
请参阅图5，图5为本申请实施例提供的后台应用程序管控装置的结构示意图。其中该后台应用程序管控装置300应用于电子设备，该后台应用程序管控装置300包括获取单元301、训练单元302和管控单元303，其中：
获取单元301，用于获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续使用状态形成第一二值向量。
具体的，预设时间段可以根据使用周期来设置，如早上8点到12点，晚上6点到10点，也可以为周一至周五，周六至周日。预设时间段可以为几个小时或几天，甚至更长的时间。然后在该预设时间段内获取预设后台应用程序的多个连续使用状态，使用状态为该预设后台应用程序是否在使用，如果在使用则记录为1，没有使用则记录为0。并且使用状态为连续的，如一个小时内，每隔5分钟记录一次使用状态，则可以获取12-13个连续使 用状态，使用状态的判断标准可以为，5分钟内只要被用户调用一次即可认为在使用。当然间隔时间可以设置为其他值，如10分钟、3分钟等。预设后台应用程序可以为网页浏览的应用、视频播放的应用、新闻浏览的应用等。
第一二值向量的数据为二进制数据的集合，如{0,1，1,0,0,0,1}，其中的数据0标示该时间段内预设后台应用程序未被调用，数据1标示该时间段内预设后台应用程序至少被调用1次。
训练单元302，用于获取参考模型，并将第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数。
其中，参考模型可以为循环神经网络模型。当然参考模型可以为混合神经网络模型、高斯混合模型或卷积神经网络模型等。
本文所使用的术语“单元”可看做为在该运算系统上执行的软件对象。本文所述的单元可看做为在该运算系统上的实施对象。而本文所述的装置及方法可以以软件的方式进行实施，当然也可在硬件上进行实施，均在本申请保护范围之内。
请一并参阅图6，图6为本申请实施例提供的后台应用程序管控装置的另一结构示意图。其中循环神经网络模型包括网络子单元3021、分类器3022、损失值计算器3023和优化子单元3024。其中：
网络子单元3021，用于将第一二值向量作为训练数据输入网络子单元得到中间值。
网络子单元3021具体包括输入层、隐层和输出层。将第一二值向量作为训练数据输入网络子单元的输入层，输入层将第一二值向量输入隐层，隐层将第一二值向量循环计算得到中间值，将中间值通过输出层输出。
分类器3022，用于将中间值输入分类器得到对应多个预测结果的概率。
在一些实施方式中，分类器3022可以基于第一预设公式将中间值输入分类器得到对应多个预测结果的概率，其中第一预设公式为：
Figure PCTCN2018102187-appb-000012
其中，Z K为目标中间值，C为预测结果的类别数，真实值可以为0或1，Z j为第j个中间值。目标中间值为第k个的中间值。预测结果根据需要设置，如后台应用程序的预测结果包括清理和不清理两类，则预测结果的类别数C为2。
损失值计算器3023，用于根据多个预测结果和与其对应的多个概率得到损失值。
在一些实施方式中，损失值计算器3023可以基于第四预设公式根据多个预测结果和与其对应的多个概率得到损失值，其中第四预设公式为：
Figure PCTCN2018102187-appb-000013
其中C为预测结果的类别数，y k为真实值，真实值可以为0或1，
Figure PCTCN2018102187-appb-000014
为第一预设公式的结果。
在一些实施方式中，损失值计算器3023可以基于第二预设公式根据多个预测结果和与其对应的多个概率得到损失值，其中第二预设公式为：
Figure PCTCN2018102187-appb-000015
其中C为预测结果的类别数，y k为真实值，
Figure PCTCN2018102187-appb-000016
为第一预设公式的结果。E为求平均值，可以获取多个第四预设公式获取的损失值，然后求平均值获取平均后的损失值，数据更准确。
需要说明的是，获取损失值除了上述方法外，还可以根据交叉熵等方式获取。
优化子单元3024，用于根据损失值进行训练，得到优化参数。
在一些实施方式中，优化子单元3024可以根据损失值利用随机梯度下降法进行训练。还可以根据批量梯度下降法或梯度下降法进行训练。
利用随机梯度下降法进行训练，可以当损失值等于或小于预设损失值时，则完成训练。也可以当连续获取的两个或多个损失值没有变化时，则完成训练。当然还可以不根据损失值，直接设定随机梯度下降法的迭代次数，迭代次数完成后，则完成训练。训练完成后，获取此时的参考模型的各个参数，并将该各个参数保存为优化参数，后续需要预测时，使用该优化参数进行预测。
其中随机梯度下降法可以采用小批量的方式训练得到最优参数。如批量大小为128。
管控单元303，用于获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续使用状态形成第二二值向量，根据参考模型、优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。
若需要判断目标后台应用程序是否可清理，首先获取该预设后台应用程序对应当前时间的多个连续使用状态，因为是后台应用程序，所以目标后台应用程序当前在后台中，当前没有被调用，获取当前时间的前几个时间段的使用状态，如以当前时刻为基准，往前每隔5分钟为一个时间段，获取前面5个连续时间段的使用状态，然后将这5个使用状态形成一个二值向量，然后输入参考模块，参考模型结合之前训练得到的优化参数，根据这个二值向量进行预测，然后生成预测结果，最后根据预测结果管控该目标后台应用程序，如关闭或保持该目标后台应用程序。
需要说明的是，参考模型的训练过程可以在服务器端也可以在电子设备端完成。当参考模型的训练过程、实际预测过程都在服务器端完成时，需要使用优化后的参考模型时，可以将预设后台应用程序当前时间前的多个时间段的使用状态输入到服务器，服务器实际预测完成后，将预测结果发送至电子设备端，电子设备再根据预测结果管控该预设后台应用程序。
当参考模型的训练过程、实际预测过程都在电子设备端完成时，需要使用优化后的参考模型时，可以将预设后台应用程序当前时间前的多个时间段的使用状态输入到电子设备，电子设备实际预测完成后，电子设备根据预测结果管控该预设后台应用程序。
请参阅图7，图7为本申请实施例提供的后台应用程序管控装置的另一应用场景示意图。当参考模型的训练过程在服务器端完成，参考模型的实际预测过程在电子设备端完成时，需要使用优化后的参考模型时，可以将预设后台应用程序当前时间前的多个时间段的 使用状态输入到电子设备，电子设备实际预测完成后，电子设备根据预测结果管控该预设后台应用程序。可选的，可以将训练好的参考模型文件(model文件)移植到智能设备上，若需要判断当前后台应用是否可清理，预设后台应用程序当前时间前的多个时间段的使用状态，输入到训练好的参考模型文件(model文件)，计算即可得到预测值。
上述所有的技术方案，可以采用任意结合形成本申请的可选实施例，在此不再一一赘述。
由上述可知，本申请实施例的后台应用程序管控装置，应用于电子设备，通过获取预设后台应用程序在预设时间段内的多个连续使用状态，并将多个连续使用状态形成第一二值向量；再将该第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；接着获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续使用状态形成第二二值向量，最后根据参考模型、所优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。可以提高对进入后台的预设后台应用程序是否还需要使用的概率进行预测的准确性，从而提升对进入后台的预测应用程序进行管控的智能化和准确性，减少内存占用。
请参阅图8，图8为本申请实施例提供的后台应用程序管控装置的又一结构示意图。本申请实施例的后台应用程序管控装置包括：评分单元304、设置单元305、获取单元301、训练单元302和管控单元303。其中：
评分单元304，用于在预设统计周期内，根据后台应用程序的运行数据对多个后台应用程序进行评分。
预设统计周期可以为一个或多个小时，一天或多天，为了获取更准确的预测结果，训练数据越多越好。
评分单元304可以基于第三预设公式根据后台应用程序的运行数据对多个后台应用程序进行评分，其中第三预设公式为：
Figure PCTCN2018102187-appb-000017
其中memory i为内存占用比，couter i为总使用次数，t i为总使用时间，seq i为预设后台应用程序的最近使用次序，Group为分组系数，G f为下沉因子，T f为时间弱化因子。memory i取值为0-1之间，seq i为预设后台应用程序的最近使用次序，例如当前使用的应用的次序值等于0，上一个使用的应用次序值为1，再之前使用的一个应用为2，依次类推。Group为分组系数，Group大于等于1，小于等于10，因为每个后台应用程序不同，用户也不同，选取的分组系数也不同。例如，可以预先计算每个后台应用程序不同分组系数的预测结果，当某个分组系数的预测结果更准确时，则采用该分组系数。G f为下沉因子，G f大于1，小于等于3，使用下沉因子后，增大了分母，进而减小了评分值，减小了seq i对评分的影响。T f为时间弱化因子，T f大于等于0，小于等于1，使用时间弱化因子后，减小了分子，进而减小了评分值，减小了t i对评分的影响。防止一个参数的权重太大，对评分有过重的影响，使各个数据之间的权重更合理。
设置单元305，用于将评分超过预设评分阈值的后台应用程序设为预设后台应用程序。
可以预先设置一个预设评分阈值，如根据评分的最高分的百分比来设置，如预设评分 阈值为评分的最高分的80％。也可以电子设备使用后，根据当前电子设备的使用情况，设定新的预设评分阈值替换之前的预设评分阈值，也可以通过网络统计一个平均值或典型值。
也可以将多个后台应用程序按照评分从高到低排列形成列表，将列表前几名的后台应用程序，如前5名，设置为预设后台应用程序。也可以将列表的某个排名的后台应用程序的评分设为预设评分阈值，如排名第6的后台应用程序的评分设为预设评分阈值。也可以不同条件下获取多个列表，然后取平均值。如每次都是取第6名的后台应用程序的评分，然后求平均值得到预设评分阈值。当然，使用一段时间后或经重新获取指令可以重新获取预设评分阈值。
考虑到时间序列分析是针对单个应用程序进行的，分析电子设备所有的应用程序数据量太大，实用性不强，因此我们要根据使用习惯选取排名较高的多个应用程序，并且清理这几个应用程序确实能大大释放内存。
本申请实施例的后台应用程序管控装置，应用于电子设备，首先选取部分后台应用程序为预设后台应用程序，然后通过获取预设后台应用程序在预设时间段内的多个连续使用状态，并将多个连续使用状态形成第一二值向量；再将该第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；接着获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续使用状态形成第二二值向量，最后根据参考模型、所优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。可以提高对进入后台的预设后台应用程序是否还需要使用的概率进行预测的准确性，从而提升对进入后台的预测应用程序进行管控的智能化和准确性，减少内存占用。
上述所有的技术方案，可以采用任意结合形成本申请的可选实施例，在此不再一一赘述。
由上述可知，本申请实施例的后台应用程序管控装置，应用于电子设备，通过获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续使用状态形成第一二值向量；将第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；将对应当前时时间的多个连续使用状态形成第二二值向量，根据参考模型、优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。可以提高对预设后台应用程序进行预测的准确性，从而提升对进入后台的应用程序进行管控的准确性。
本申请实施例中，后台应用程序管控装置与上文实施例中的后台应用程序管控方法属于同一构思，在后台应用程序管控装置上可以运行后台应用程序管控方法实施例中提供的任一方法，其具体实现过程详见后台应用程序管控方法的实施例，此处不再赘述。
本申请实施例还提供一种电子设备。请参阅图9，电子设备400包括处理器401以及存储器402。其中，处理器401与存储器402电性连接。
处理器400是电子设备400的控制中心，利用各种接口和线路连接整个电子设备的各个部分，通过运行或加载存储在存储器402内的计算机程序，以及调用存储在存储器402内的数据，执行电子设备400的各种功能并处理数据，从而对电子设备400进行整体监控。
存储器402可用于存储软件程序以及模块，处理器401通过运行存储在存储器402的 计算机程序以及模块，从而执行各种功能应用以及数据处理。存储器402可主要包括存储程序区和存储数据区，其中，存储程序区可存储操作系统、至少一个功能所需的计算机程序(比如声音播放功能、图像播放功能等)等；存储数据区可存储根据电子设备的使用所创建的数据等。此外，存储器402可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他易失性固态存储器件。相应地，存储器402还可以包括存储器控制器，以提供处理器401对存储器402的访问。
在本申请实施例中，电子设备400中的处理器401会按照如下的流程，将一个或一个以上的计算机程序的进程对应的指令加载到存储器402中，并由处理器401运行存储在存储器402中的计算机程序，从而实现各种功能，如下：
获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续使用状态形成第一二值向量；获取参考模型，并将第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续使用状态形成第二二值向量，根据参考模型、优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。
在一些实施方式中，处理器401还用于执行：
选取循环神经网络模型为参考模型。
在一些实施方式中，处理器401还用于执行：
循环神经网络模型包括网络子单元和分类器；
将第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数具体包括：
将第一二值向量作为训练数据输入网络子单元得到中间值；
将中间值输入分类器得到对应多个预测结果的概率；
根据多个预测结果和与其对应的多个概率得到损失值；
根据损失值进行训练，得到优化参数。
在一些实施方式中，处理器401还用于执行：
基于第一预设公式将中间值输入分类器得到对应多个预测结果的概率，其中第一预设公式为：
Figure PCTCN2018102187-appb-000018
其中，Z K为目标中间值，C为预测结果的类别数，Z j为第j个中间值。
在一些实施方式中，处理器401还用于执行：
基于第二预设公式根据多个预测结果和与其对应的多个概率得到损失值，其中第二预设公式为：
Figure PCTCN2018102187-appb-000019
其中y k为真实值，E为求平均值。
在一些实施方式中，处理器401还用于执行：
基于第四预设公式根据多个所述预测结果和与其对应的多个所述概率得到损失值，其中所述第四预设公式为：
Figure PCTCN2018102187-appb-000020
其中
Figure PCTCN2018102187-appb-000021
为第一预设公式的结果，E为求平均值。
在一些实施方式中，处理器401还用于执行：
根据损失值利用随机梯度下降法进行训练。
在一些实施方式中，处理器401还用于执行：
在预设统计周期内，根据后台应用程序的运行数据对多个后台应用程序进行评分；
将评分超过预设评分阈值的后台应用程序设为预设后台应用程序。
在一些实施方式中，处理器401还用于执行：
基于第三预设公式根据后台应用程序的运行数据对多个后台应用程序进行评分，其中第三预设公式为：
Figure PCTCN2018102187-appb-000022
其中memory i为内存占用比，couter i为总使用次数，t i为总使用时间，seq i为预设后台应用程序的最近使用次序，Group为分组系数， Gf为下沉因子，T f为时间弱化因子。
由上述可知，本申请实施例提供的电子设备，通过获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续使用状态形成第一二值向量；将第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；将对应当前时时间的多个连续使用状态形成第二二值向量，根据参考模型、优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。可以提高对预设后台应用程序进行预测的准确性，从而提升对进入后台的应用程序进行管控的准确性。
请一并参阅图10，在一些实施方式中，电子设备400还可以包括：显示器403、射频电路404、音频电路405以及电源406。其中，其中，显示器403、射频电路404、音频电路405以及电源406分别与处理器401电性连接。
显示器403可以用于显示由用户输入的信息或提供给用户的信息以及各种图形用户接口，这些图形用户接口可以由图形、文本、图标、视频和其任意组合来构成。显示器403可以包括显示面板，在一些实施方式中，可以采用液晶显示器(Liquid Crystal Display，LCD)、或者有机发光二极管(Organic Light-Emitting Diode，OLED)等形式来配置显示面板。
射频电路404可以用于收发射频信号，以通过无线通信与网络设备或其他电子设备建 立无线通讯，与网络设备或其他电子设备之间收发信号。
音频电路405可以用于通过扬声器、传声器提供用户与电子设备之间的音频接口。
电源406可以用于给电子设备400的各个部件供电。在一些实施方式中，电源406可以通过电源管理系统与处理器401逻辑相连，从而通过电源管理系统实现管理充电、放电、以及功耗管理等功能。
尽管图10中未示出，电子设备400还可以包括摄像头、蓝牙模块等，在此不再赘述。
本申请实施例还提供一种存储介质，存储介质存储有计算机程序，当计算机程序在计算机上运行时，使得计算机执行上述任一实施例中的应用程序管控方法，比如：获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续使用状态形成第一二值向量；获取参考模型，并将第一二值向量作为训练数据输入参考模型进行训练，得到训练后的参考模型的优化参数；获取预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续使用状态形成第二二值向量，根据参考模型、优化参数以及第二二值向量进行预测，生成预测结果，并根据预测结果对预设后台应用程序进行管控。
在本申请实施例中，存储介质可以是磁碟、光盘、只读存储器(Read Only Memory，ROM)、或者随机存取记忆体(Random Access Memory，RAM)等。
在上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述的部分，可以参见其他实施例的相关描述。
需要说明的是，对本申请实施例的后台应用程序管控方法而言，本领域普通测试人员可以理解实现本申请实施例后台应用程序管控方法的全部或部分流程，是可以通过计算机程序来控制相关的硬件来完成，计算机程序可存储于一计算机可读取存储介质中，如存储在电子设备的存储器中，并被该电子设备内的至少一个处理器执行，在执行过程中可包括如后台应用程序管控方法的实施例的流程。其中，的存储介质可为磁碟、光盘、只读存储器、随机存取记忆体等。
对本申请实施例的后台应用程序管控装置而言，其各功能模块可以集成在一个处理芯片中，也可以是各个模块单独物理存在，也可以两个或两个以上模块集成在一个模块中。上述集成的模块既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。集成的模块如果以软件功能模块的形式实现并作为独立的产品销售或使用时，也可以存储在一个计算机可读取存储介质中，存储介质譬如为只读存储器，磁盘或光盘等。
以上对本申请实施例所提供的一种后台应用程序管控方法、装置、存储介质及电子设备进行了详细介绍，本文中应用了具体个例对本申请的原理及实施方式进行了阐述，以上实施例的说明只是用于帮助理解本申请的方法及其核心思想；同时，对于本领域的技术人员，依据本申请的思想，在具体实施方式及应用范围上均会有改变之处，综上所述，本说明书内容不应理解为对本申请的限制。

Claims (20)

  1. 一种后台应用程序管控方法，应用于电子设备，所述方法包括：
    获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续所述使用状态形成第一二值向量；
    获取参考模型，并将所述第一二值向量作为训练数据输入所述参考模型进行训练，得到训练后的所述参考模型的优化参数；
    获取所述预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续所述使用状态形成第二二值向量，根据所述参考模型、所述优化参数以及所述第二二值向量进行预测，生成预测结果，并根据所述预测结果对所述预设后台应用程序进行管控。
  2. 根据权利要求1所述的后台应用程序管控方法，其中，所述参考模型为循环神经网络模型。
  3. 根据权利要求2所述的后台应用程序管控方法，其中，所述循环神经网络模型包括网络子单元和分类器；
    所述将所述第一二值向量作为训练数据输入所述参考模型进行训练，得到训练后的所述参考模型的优化参数具体包括：
    将第一二值向量作为训练数据输入所述网络子单元得到中间值；
    将所述中间值输入所述分类器得到对应多个所述预测结果的概率；
    根据多个所述预测结果和与其对应的多个所述概率得到损失值；
    根据所述损失值进行训练，得到所述优化参数。
  4. 根据权利要求3所述的后台应用程序管控方法，其中，所述将所述中间值输入所述分类器得到对应多个所述预测结果的概率具体包括：
    基于第一预设公式将所述中间值输入所述分类器得到对应多个所述预测结果的概率，其中所述第一预设公式为：
    Figure PCTCN2018102187-appb-100001
    其中，Z K为目标中间值，C为预测结果的类别数，Z j为第j个中间值。
  5. 根据权利要求4所述的后台应用程序管控方法，其中，所述根据多个所述预测结果和与其对应的多个所述概率得到损失值具体包括：
    基于第二预设公式根据多个所述预测结果和与其对应的多个所述概率得到损失值，其中所述第二预设公式为：
    Figure PCTCN2018102187-appb-100002
    其中y k为真实值，E为求平均值。
  6. 根据权利要求4所述的后台应用程序管控方法，其中，所述根据多个所述预测结果 和与其对应的多个所述概率得到损失值具体包括：
    基于第四预设公式根据多个所述预测结果和与其对应的多个所述概率得到损失值，其中所述第四预设公式为：
    Figure PCTCN2018102187-appb-100003
    其中
    Figure PCTCN2018102187-appb-100004
    为第一预设公式的结果，E为求平均值。
  7. 根据权利要求4所述的后台应用程序管控方法，其中，所述根据多个所述预测结果和与其对应的多个所述概率得到损失值具体包括：
    基于交叉熵模型，根据多个所述预测结果和与其对应的多个所述概率得到损失值。
  8. 根据权利要求3所述的后台应用程序管控方法，其中，所述根据所述损失值进行训练具体包括：
    根据所述损失值利用随机梯度下降法或批量梯度下降法进行训练。
  9. 根据权利要求1所述的后台应用程序管控方法，其中，所述获取预设后台应用程序在预设时间段内的多个连续使用状态之前，所述方法还包括：
    在预设统计周期内，根据后台应用程序的运行数据对多个后台应用程序进行评分；
    将评分超过预设评分阈值的后台应用程序设为所述预设后台应用程序。
  10. 根据权利要求9所述的后台应用程序管控方法，其中，所述根据后台应用程序的运行数据对多个后台应用程序进行评分具体包括：
    基于第三预设公式根据后台应用程序的运行数据对多个后台应用程序进行评分，其中所述第三预设公式为：
    Figure PCTCN2018102187-appb-100005
    其中memory i为内存占用比，couter i为总使用次数，t i为总使用时间，seq i为预设后台应用程序的最近使用次序，Group为分组系数，G f为下沉因子，T f为时间弱化因子。
  11. 根据权利要求1所述的后台应用程序管控方法，其中，根据所述预测结果对所述预设后台应用程序进行管控具体包括：
    当预测结果为关闭时，将所述预设后台应用程序关闭；
    当预测结果为保留时，将所述预设后台应用程序保留。
  12. 一种存储介质，其上存储有计算机程序，其中，当所述计算机程序在计算机上运行时，使得所述计算机执行如权利要求1至11任一项所述的后台应用程序管控方法。
  13. 一种电子设备，包括处理器和存储器，所述存储器有计算机程序，所述处理器调用所述计算机程序，其中，所述处理器还执行：
    获取预设后台应用程序在预设时间段内的多个连续使用状态，并将获取的多个连续所述使用状态形成第一二值向量；
    获取参考模型，并将所述第一二值向量作为训练数据输入所述参考模型进行训练，得到训练后的所述参考模型的优化参数；
    获取所述预设后台应用程序对应当前时间的多个连续使用状态，并将对应当前时时间的多个连续所述使用状态形成第二二值向量，根据所述参考模型、所述优化参数以及所述第二二值向量进行预测，生成预测结果，并根据所述预测结果对所述预设后台应用程序进行管控。
  14. 根据权利要求13所述的电子设备，其中，所述参考模型为循环神经网络模型。
  15. 根据权利要求14所述的电子设备，其中，所述循环神经网络模型包括网络子单元和分类器；
    在将所述第一二值向量作为训练数据输入所述参考模型进行训练，得到训练后的所述参考模型的优化参数中，所述处理器还执行：
    将第一二值向量作为训练数据输入所述网络子单元得到中间值；
    将所述中间值输入所述分类器得到对应多个所述预测结果的概率；
    根据多个所述预测结果和与其对应的多个所述概率得到损失值；
    根据所述损失值进行训练，得到所述优化参数。
  16. 根据权利要求15所述的电子设备，其中，在将所述中间值输入所述分类器得到对应多个所述预测结果的概率中，所述处理器还执行：
    基于第一预设公式将所述中间值输入所述分类器得到对应多个所述预测结果的概率，其中所述第一预设公式为：
    Figure PCTCN2018102187-appb-100006
    其中，Z K为目标中间值，C为预测结果的类别数，Z j为第j个中间值。
  17. 根据权利要求16所述的电子设备，其中，在根据多个所述预测结果和与其对应的多个所述概率得到损失值中，所述处理器还执行：
    基于第二预设公式根据多个所述预测结果和与其对应的多个所述概率得到损失值，其中所述第二预设公式为：
    Figure PCTCN2018102187-appb-100007
    其中y k为真实值，E为求平均值。
  18. 根据权利要求16所述的电子设备，其中，在根据多个所述预测结果和与其对应的多个所述概率得到损失值中，所述处理器还执行：
    基于第四预设公式根据多个所述预测结果和与其对应的多个所述概率得到损失值，其中所述第四预设公式为：
    Figure PCTCN2018102187-appb-100008
    其中
    Figure PCTCN2018102187-appb-100009
    为第一预设公式的结果，E为求平均值。
  19. 根据权利要求13所述的电子设备，其中，在获取预设后台应用程序在预设时间段 内的多个连续使用状态之前，所述处理器还执行：
    在预设统计周期内，根据后台应用程序的运行数据对多个后台应用程序进行评分；
    将评分超过预设评分阈值的后台应用程序设为所述预设后台应用程序。
  20. 根据权利要求19所述的电子设备，其中，在根据后台应用程序的运行数据对多个后台应用程序进行评分中，所述处理器还执行：
    基于第三预设公式根据后台应用程序的运行数据对多个后台应用程序进行评分，其中所述第三预设公式为：
    Figure PCTCN2018102187-appb-100010
    其中memory i为内存占用比，couter i为总使用次数，t i为总使用时间，seq i为预设后台应用程序的最近使用次序，Group为分组系数，G f为下沉因子，T f为时间弱化因子。
PCT/CN2018/102187 2017-09-30 2018-08-24 后台应用程序管控方法、存储介质及电子设备 WO2019062409A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710922671.7A CN107704364B (zh) 2017-09-30 2017-09-30 后台应用程序管控方法、装置、存储介质及电子设备
CN201710922671.7 2017-09-30

Publications (1)

Publication Number Publication Date
WO2019062409A1 true WO2019062409A1 (zh) 2019-04-04

Family

ID=61184018

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/102187 WO2019062409A1 (zh) 2017-09-30 2018-08-24 后台应用程序管控方法、存储介质及电子设备

Country Status (2)

Country Link
CN (1) CN107704364B (zh)
WO (1) WO2019062409A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107678845B (zh) * 2017-09-30 2020-03-10 Oppo广东移动通信有限公司 应用程序管控方法、装置、存储介质及电子设备
CN107704364B (zh) * 2017-09-30 2019-08-06 Oppo广东移动通信有限公司 后台应用程序管控方法、装置、存储介质及电子设备
CN108984369A (zh) * 2018-07-13 2018-12-11 厦门美图移动科技有限公司 卡顿预测方法、装置及移动终端

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103218566A (zh) * 2013-01-25 2013-07-24 江南大学 基于Android平台软件行为检测的主动防御系统
CN103902474A (zh) * 2014-04-11 2014-07-02 华中科技大学 一种支持固态盘缓存动态分配的混合存储系统和方法
CN104239194A (zh) * 2014-09-12 2014-12-24 上海交通大学 基于bp神经网络的任务完成时间预测方法
CN106294902A (zh) * 2015-05-28 2017-01-04 阿里巴巴集团控股有限公司 预测移动应用程序页面性能的方法、装置及电子设备
CN107704364A (zh) * 2017-09-30 2018-02-16 广东欧珀移动通信有限公司 后台应用程序管控方法、装置、存储介质及电子设备

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102521041B (zh) * 2011-12-14 2014-10-08 华为终端有限公司 一种处理应用程序的方法及无线手持设备
CN105718027B (zh) * 2016-01-20 2019-05-31 努比亚技术有限公司 后台应用程序的管理方法及移动终端
CN105843650B (zh) * 2016-03-31 2020-03-31 青岛海信移动通信技术股份有限公司 一种智能终端中的应用程序管理方法和装置
CN107133094B (zh) * 2017-06-05 2021-11-02 努比亚技术有限公司 应用管理方法、移动终端及计算机可读存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103218566A (zh) * 2013-01-25 2013-07-24 江南大学 基于Android平台软件行为检测的主动防御系统
CN103902474A (zh) * 2014-04-11 2014-07-02 华中科技大学 一种支持固态盘缓存动态分配的混合存储系统和方法
CN104239194A (zh) * 2014-09-12 2014-12-24 上海交通大学 基于bp神经网络的任务完成时间预测方法
CN106294902A (zh) * 2015-05-28 2017-01-04 阿里巴巴集团控股有限公司 预测移动应用程序页面性能的方法、装置及电子设备
CN107704364A (zh) * 2017-09-30 2018-02-16 广东欧珀移动通信有限公司 后台应用程序管控方法、装置、存储介质及电子设备

Also Published As

Publication number Publication date
CN107704364B (zh) 2019-08-06
CN107704364A (zh) 2018-02-16

Similar Documents

Publication Publication Date Title
US10579910B2 (en) Dynamic classifier selection based on class skew
WO2019062417A1 (zh) 应用清理方法、装置、存储介质及电子设备
US20200225995A1 (en) Application cleaning method, storage medium and electronic device
WO2019062414A1 (zh) 应用程序管控方法、装置、存储介质及电子设备
WO2019120019A1 (zh) 用户性别预测方法、装置、存储介质及电子设备
CN111045814B (zh) 资源调度方法和终端设备
WO2017140221A1 (zh) 文本信息处理方法和装置
WO2019062413A1 (zh) 应用程序管控方法、装置、存储介质及电子设备
WO2019062342A9 (zh) 后台应用清理方法、装置、存储介质及电子设备
CN107835311B (zh) 应用管理方法、装置、存储介质及电子设备
US20180367325A1 (en) Method and system for sorting chatroom list based on conversational activeness and contextual information
CN107870810B (zh) 应用清理方法、装置、存储介质及电子设备
WO2019062409A1 (zh) 后台应用程序管控方法、存储介质及电子设备
WO2019062416A1 (zh) 应用清理方法、装置、存储介质及电子设备
WO2019062405A1 (zh) 应用程序的处理方法、装置、存储介质及电子设备
WO2019085754A1 (zh) 应用清理方法、装置、存储介质及电子设备
WO2019062419A1 (zh) 应用清理方法、装置、存储介质及电子设备
WO2019120007A1 (zh) 用户性别预测方法、装置及电子设备
US20220179706A1 (en) Adaptive resource allocation system and method for a target application executed in an information handling system (ihs)
WO2019062341A1 (zh) 后台应用清理方法、装置、存储介质及电子设备
WO2019062404A1 (zh) 应用程序的处理方法、装置、存储介质及电子设备
WO2019062411A1 (zh) 后台应用程序管控方法、存储介质及电子设备
CN107943537B (zh) 应用清理方法、装置、存储介质及电子设备
CN107943535B (zh) 应用清理方法、装置、存储介质及电子设备
US20230214706A1 (en) Automated generation of agent configurations for reinforcement learning

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18860657

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18860657

Country of ref document: EP

Kind code of ref document: A1