WO2019183781A1 - 一种数据处理方法及网络设备 - Google Patents

一种数据处理方法及网络设备 Download PDF

Info

Publication number
WO2019183781A1
WO2019183781A1 PCT/CN2018/080568 CN2018080568W WO2019183781A1 WO 2019183781 A1 WO2019183781 A1 WO 2019183781A1 CN 2018080568 W CN2018080568 W CN 2018080568W WO 2019183781 A1 WO2019183781 A1 WO 2019183781A1
Authority
WO
WIPO (PCT)
Prior art keywords
power consumption
application
threshold
mobile devices
data
Prior art date
Application number
PCT/CN2018/080568
Other languages
English (en)
French (fr)
Inventor
张�浩
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Priority to PCT/CN2018/080568 priority Critical patent/WO2019183781A1/zh
Priority to CN201880090873.2A priority patent/CN111819550B/zh
Publication of WO2019183781A1 publication Critical patent/WO2019183781A1/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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • the embodiment of the invention relates to the field of data processing, and in particular to a data processing method and a network device.
  • the Android system brings many cheaper users, such as easy to use and low cost, but it also brings many problems, such as the power consumption of third-party applications. Due to the variety of third-party applications, the power consumption of the Android system is large. At present, the Android system mainly monitors the power consumption of third-party applications through the background top application, but since there is no strategy for comprehensive monitoring of the power consumption of the mobile device, the system power consumption of the mobile device cannot be effectively controlled.
  • the embodiment of the invention discloses a data processing method and a network device, which are used for providing a comprehensive monitoring strategy for power consumption of a mobile device, so that the power consumption of the mobile device can be effectively controlled and controlled.
  • the first aspect discloses a data processing method, which is applied to a network device, acquires first data including data of M mobile devices, determines predicted power consumption according to the first data, and calculates actual power consumption and predicted power consumption according to the M mobile devices. Determining a first threshold, transmitting the predicted power consumption and the first threshold to the M mobile devices, the predicted power consumption and the first threshold are used to indicate that the first mobile device monitors actual power consumption of the first mobile device, when the first mobile device When the difference between the actual power consumption and the predicted power consumption is greater than the first threshold, determining that the total power consumption of the first mobile device is in an abnormal state, so that the power consumption of the M mobile devices can be performed according to the first threshold and the predicted power consumption. Effective control.
  • the data of the first mobile device includes power consumption and duration of the first mobile device using the device, and the power consumption of the first mobile device, the duration of the bright screen, and the duration of the screen blanking.
  • the first mobile device is any of the M mobile devices.
  • a mobile device, M is an integer greater than two.
  • the first feature vector when determining the predicted power consumption according to the first data, may be first determined according to the first data, and then the predicted power consumption is determined according to the first feature vector and the linear regression model, so that the predicted power can be quickly determined. Consumption.
  • the difference between the actual power consumption and the predicted power consumption of each of the M mobile devices may be calculated first. Obtaining a power consumption difference, and then performing normal distribution processing on the power consumption difference to obtain a normal distribution difference value, and determining a first threshold value from the normal distribution difference value according to a preset rule, so that a first precision is obtained. Threshold.
  • the second data of the data of the N mobile devices is acquired, the applications on the N mobile devices are classified according to the second data to obtain the H-type application, and the threshold of each type of application in the H-type application is determined.
  • the threshold of each type of application in the H-type application is sent to the N mobile devices, and the second threshold is used to indicate that the second mobile device monitors the total background power consumption of the first application, when the total background power consumption of the first application is greater than the second threshold.
  • the power consumption of the first application is determined to be in an abnormal state, so that different types of applications can be effectively controlled according to the threshold of each type of application in the H-type application, and the threshold of each application can be dynamically adjusted.
  • the data of the second mobile device includes the total background power consumption of the application device on the second mobile device, and the second mobile device is any one of the N mobile devices, where N is an integer greater than 2, and H is greater than 2.
  • the first application is any application belonging to the H-type application on the second mobile device, and the second threshold is a threshold of the application class to which the first application belongs.
  • the cumulative distribution function of the total background power consumption of the applications on the N mobile devices may be first drawn according to the second data ( Cumulative Distribution Function (CDF) curve, the number of categories H applied on N mobile devices is determined according to the CDF curve, and the applications on N mobile devices are classified into H classes according to the CDF curve and the Convolutional Neural Network (CNN).
  • CDF Cumulative Distribution Function
  • CNN Convolutional Neural Network
  • determining a threshold for each type of application in the H-type application that is, determining a total background power consumption of the second application on the L mobile devices in the N mobile devices as a threshold of the target class application, L mobile devices
  • the total background power consumption of the second application is the same, the ratio of L to N is equal to the preset ratio, the target application is any application in the H application, and the second application is the application with the largest power consumption in the target application. .
  • the third data of the data of the K mobile devices is acquired, the thresholds applied on the K mobile devices are determined according to the third data, and the thresholds applied by the K mobile devices are sent to the K mobile devices.
  • the third threshold is used to indicate that the third mobile device monitors power consumption of the third application, and when the difference between the power consumption of the third application and the third threshold is greater than a preset value, determining that the power consumption of the third application is in an abnormal state, so as to be
  • the thresholds applied on K mobile devices are systematically and effectively controlled for each application.
  • the data of the third mobile device includes the duration of the application using the device on the third mobile device, the third mobile device is any one of the K mobile devices, K is an integer greater than 2, and the third application is the third mobile The application on the device, the third threshold is the threshold of the third application.
  • the second feature vector when determining the threshold applied on the K mobile devices according to the third data, may be first determined according to the third data, and then according to the second feature vector and the Support Vector Machine (SVM).
  • SVM Support Vector Machine
  • the thresholds applied on the K mobile devices are determined, or the thresholds applied on the K mobile devices are determined according to the second feature vector and the Deep Neural Network (DNN) model.
  • DNN Deep Neural Network
  • a second aspect discloses a network device comprising means for performing the data processing method provided by the first aspect or any of the possible implementations of the first aspect.
  • a third aspect discloses a network device comprising a processor, a memory and a transceiver, a memory for storing program code, a processor for executing program code, and a transceiver for communicating with the mobile device.
  • the processor executes the program code stored in the memory, the processor is caused to perform the data processing method disclosed in any of the possible implementations of the first aspect or the first aspect.
  • a fourth aspect discloses a readable storage medium storing program code for a network device to perform the data processing method disclosed in the first aspect or any of the possible implementations of the first aspect.
  • FIG. 1 is a schematic flow chart of a data processing method according to an embodiment of the present invention.
  • FIG. 3 is a normal distribution diagram of a power consumption difference disclosed in an embodiment of the present invention.
  • FIG. 4 is a schematic diagram of a box pattern principle disclosed in an embodiment of the present invention.
  • FIG. 5 is a schematic diagram of an abnormal value determined by a box pattern according to an embodiment of the present invention.
  • FIG. 6 is a schematic diagram of a power consumption change with time according to an embodiment of the present invention.
  • FIG. 7 is a schematic flowchart diagram of another data processing method according to an embodiment of the present disclosure.
  • FIG. 8 is a schematic diagram of a CDF curve disclosed in an embodiment of the present invention.
  • FIG. 10 is a normal distribution diagram of a CDF area disclosed in an embodiment of the present invention.
  • FIG. 11 is a schematic diagram of a contour coefficient disclosed in an embodiment of the present invention.
  • FIG. 12 is a schematic diagram of a CDF curve of a classified application disclosed in an embodiment of the present invention.
  • FIG. 13 is a schematic structural diagram of a network device according to an embodiment of the present disclosure.
  • FIG. 14 is a schematic structural diagram of another network device according to an embodiment of the present disclosure.
  • FIG. 15 is a schematic structural diagram of still another network device according to an embodiment of the present invention.
  • the embodiment of the invention discloses a data processing method and a network device, which are used for providing a comprehensive monitoring strategy for power consumption of a mobile device, so that the power consumption of the mobile device can be effectively controlled and controlled. The details are described below.
  • FIG. 1 is a schematic flowchart diagram of a data processing method according to an embodiment of the present invention.
  • the data processing method is described from the perspective of a network device.
  • the data processing method may include the following steps.
  • the network device may periodically acquire the first data, where the first data includes data of the M mobile devices, and may be actively acquired by the network device, or may be actively sent by the M mobile devices.
  • the data of the first mobile device includes the power consumption and duration of the first mobile device, and the power consumption of the first mobile device, the duration of the bright screen, and the duration of the screen.
  • the period of data collected by the first mobile device may be one hour. , two hours, one day, etc.
  • the power consumption and duration of the first mobile device using the device may be the power consumption and duration of all devices used by the first mobile device; or the power consumption and duration of the first mobile device using all devices; or may be the first
  • the power consumption and duration of the A application that consumes the most power in all applications on the mobile device A is an integer greater than or equal to 2; it can also be the work of the first mobile device using the most power-consuming B devices in all devices.
  • Consumption and duration, B is an integer greater than or equal to 2; it can also be the power consumption and duration of B devices in all devices using the most power-consuming A applications in all applications on the first mobile device.
  • the first data is the power consumption of 12 of the 10 most power-consuming applications (DISPLAY, CPU, GNSS, SENSOR, GPU, FRONTCAM, REARCAM, FLASHLIGHT, AUDIO, MODEM, WIFI, BT). duration.
  • the power consumption and duration of the device are the total power consumption of the front and back of the device and the total duration of use of the front and back.
  • the first mobile device is any one of the M mobile devices that establish a connection with the network device, and M is an integer greater than 2.
  • the first data may be optimized, that is, the abnormal data in the first data is filtered, that is, the noise in the first data is filtered out.
  • the optimization process of the first data includes: the statistical duration is equal to the sum of the charging duration and the non-charging duration, or equal to the sum of the duration of the screen and the duration of the bright screen; the statistical duration is greater than Or equal to 22 hours and less than or equal to 26 hours; the first mobile device's daily power consumption is the sum of the bright screen power consumption and the screen power consumption divided by the battery capacity and multiplied by 100; the ratio of the charging duration to the statistical duration is less than or It is equal to 0.3333; the power consumption of the first mobile device is greater than or equal to 5 grids per day, wherein the total amount of power of the first mobile device is divided into 100 grids.
  • the network device determines the predicted power consumption according to the first data, that is, determines the first feature vector according to the first data, and determines the predicted power consumption according to the first feature vector and the linear regression model.
  • the first data herein may be the first data directly obtained by the network device, or may be the first data after being optimized.
  • the dimension of the feature vector may be determined first.
  • the systems of the M mobile devices are version 126, and the period of data collected by the M mobile devices is one day, it is assumed that the power consumption and duration of the devices used by the M mobile devices per day are used for each application on a daily basis for each mobile device.
  • the top 100 power consumption applications use the power consumption and duration of 12 of all devices.
  • the 12 devices are DISPLAY, CPU, GNSS, SENSOR, GPU, FRONTCAM, REARCAM, FLASHLIGHT, AUDIO, MODEM, WIFI, BT. .
  • the vector dimension is 2830 (100*14*2+14*2+2), wherein 100 of 100*14*2 is the number of applications, 14 is the number of devices, and since MODEM and WIFI are divided into uplink and downlink, The number of devices is not 12, but 14,2 indicates that the data is collected in the front and back; 14*2 is used to characterize the data dimensions of the twelve devices used by applications other than the above 100 applications, 14 and 2 The meaning is the same as before; the last 2 means the duration of the screen is off and off.
  • the third feature vector is determined according to the first data and the dimension of the feature vector. Among them, the data of unused devices is zero-padded to ensure that the dimensions of each vector are the same. Then, the third feature vector is normalized or normalized to obtain a fourth feature vector, so as to de-scale the data in the third feature vector, thereby unifying the unit-unit data in the third feature vector into unitless data. Then, the feature with less correlation in the fourth feature vector is removed to obtain the first feature vector, that is, when the feature correlation is small, the feature is deleted.
  • the predicted power consumption is then determined based on the first eigenvector and the linear regression model, ie, the first eigenvector is used as an input to the linear regression model to determine the predicted power consumption.
  • the results under different parameters can be tried, and the result of the linear regression model determined by selecting the parameter with the smallest error is the predicted power consumption.
  • the first threshold is determined according to the actual power consumption and the predicted power consumption of the M mobile devices.
  • the difference between the actual power consumption and the predicted power consumption of each of the M mobile devices may be calculated to obtain a power consumption difference value, and then the power consumption difference value is normally distributed to obtain a normal distribution difference value.
  • FIG. 2 is a histogram of the power consumption difference disclosed in the embodiment of the present invention. As shown in FIG. 2, the abscissa is the power consumption difference dv, and the ordinate is the frequency at which the power consumption difference occurs.
  • FIG. 3 is a normal distribution diagram of a power consumption difference disclosed in an embodiment of the present invention. As shown in FIG. 3, FIG. 3 is a normal distribution diagram after the histogram of the power consumption difference shown in FIG. 2 is normally distributed. Then, the first threshold is determined from the normal distribution difference according to a preset rule, and the difference at ⁇ +3 ⁇ in FIG. 3 can be determined as the first threshold according to the 3 ⁇ criterion in statistics.
  • the first threshold is determined by using a box diagram.
  • FIG. 4 is a schematic diagram of a box diagram principle disclosed in an embodiment of the present invention.
  • the lower quartile is Q1
  • the median is Q2
  • the power consumption difference is in the [Q1-3IQR, Q3+3IQR] interval.
  • the value other than the value is determined as the severe abnormal value, and the value whose power consumption difference is outside the interval of [Q1-1.5IQR, Q3+1.5IQR] is determined as the mild abnormal value.
  • FIG. 5 is a schematic diagram of an abnormal value determined by a box pattern according to an embodiment of the present invention. As shown in Fig.
  • FIG. 6 is a schematic diagram of power consumption changing with time according to an embodiment of the present invention.
  • the abscissa is time (in seconds) and the ordinate is power consumption (in mAh). It can be seen that as time increases, the power consumption of mobile devices increases, and the total power consumption of mobile devices The abnormal state is also getting worse.
  • the predicted power consumption and the first threshold are respectively sent to the M mobile devices, so that the first mobile device monitors the first The actual power consumption of the mobile device, when the difference between the actual power consumption of the first mobile device and the predicted power consumption is greater than the first threshold, determining that the total power consumption of the first mobile device is in an abnormal state, and the first mobile device may be frozen, Kill or restart some or all of the background apps on the first mobile device.
  • FIG. 7 is a schematic flowchart diagram of another data processing method according to an embodiment of the present invention.
  • the data processing method is described from the perspective of a network device. As shown in FIG. 7, the data processing method may include the following steps.
  • Step 701 is the same as step 101. For details, refer to step 101, and details are not described herein.
  • Step 702 is the same as step 102.
  • Step 102 For detailed description, refer to step 102, and details are not described herein again.
  • Step 703 is the same as step 103.
  • Step 703 For details, refer to step 103, and details are not described herein.
  • Step 704 is the same as step 104.
  • Step 104 For detailed description, refer to step 104, and details are not described herein.
  • the network device may periodically acquire the second data, where the second data includes data of the N mobile devices, which may be actively acquired by the network device, or may be actively sent by the N mobile devices.
  • the data of the second mobile device includes the total background power consumption of the application device on the second mobile device, and the period of the second mobile device data collection may be one hour, two hours, one day, and the like.
  • the total background power consumption of the application device on the second mobile device may be the total background power consumption of all devices on the second mobile device using all devices, or may be the highest power consumption of all applications on the second mobile device.
  • the application uses the total background power consumption of the device; it can also be the total background power consumption of the B devices in all devices using the highest power consumption of all applications in the second mobile device.
  • the second data is the background total work of 12 of the 10 most power-consuming applications (DISPLAY, CPU, GNSS, SENSOR, GPU, FRONTCAM, REARCAM, FLASHLIGHT, AUDIO, MODEM, WIFI, BT). Consumption.
  • the second mobile device is any one of the N mobile devices, and N is an integer greater than 2.
  • FIG. 8 is a schematic diagram of a CDF curve disclosed in an embodiment of the present invention. As shown in Figure 8, the abscissa of the CDF curve is the power consumption (in mAh) and the ordinate is the cumulative percentage (in %). Then, according to the CDF curve, the number of categories H applied on the N mobile devices is determined, and the area of the first area in FIG. 8 can be calculated first to obtain the CDF area. Please refer to FIG. 9. FIG.
  • FIG. 9 is a histogram of a CDF area disclosed in an embodiment of the present invention.
  • the abscissa is the CDF area area
  • the ordinate is the frequency at which the CDF area appears.
  • FIG. 10 is a normal distribution diagram of a CDF area according to an embodiment of the present invention.
  • Fig. 10 is a view showing a histogram of the CDF area in Fig. 9 after normal distribution processing. Then, kmean clustering is performed on the CDF area shown in FIG. 10 to obtain a contour coefficient.
  • FIG. 11 is a schematic diagram of a contour coefficient disclosed in an embodiment of the present invention. As shown in FIG.
  • the abscissa is the number of categories of the cluster, and the ordinate is the contour coefficient.
  • the number of categories corresponding to the local peaks of the contour coefficients selected from the curves shown in FIG. 11 are 7, 8, 11, 17, and 20, respectively.
  • the applications are classified according to the number of categories and the CDF area. For example, suppose the number of categories of the applied classification is 7, the CDF area is 10-101, the maximum CDF area is subtracted from the minimum CDF area to obtain 91, and then the 91 is divided by 7 to obtain 13, which can respectively set the CDF area to 10-23. Applications between, between 23-36, between 36-49, between 49-62, between 62-75, between 75-88 and between 88-101 were identified as the same type of application.
  • FIG. 12 is a schematic diagram of a CDF curve of a classified application according to an embodiment of the present invention. As shown in Figure 12, the number of categories to which the classification is applied is 10, and each type of application is the closest application to the CDF curve. Where H is an integer greater than 2.
  • the threshold of each type of application in the H-type application is determined, and L of the N mobile devices may be moved.
  • the total background power consumption of the second application on the device is determined as the threshold of the target application, and the total background power consumption of the second application on the L mobile devices is the same, and the ratio of L to N is equal to the preset ratio, and the target application is For any type of application in a class H application, the second application is the one with the highest power consumption in the target class application. As shown in FIG.
  • the horizontal line in the figure is a horizontal line corresponding to 80% of the ordinate, and the value of the abscissa corresponding to the intersection of the horizontal line and the CDF curve can be determined as the threshold value of each type of application in the H-type application. Since there is more than one intersection of each type of application in the horizontal line and the H-type application, the maximum horizontal coordinate value corresponding to the intersection point in each type of application can be determined as the threshold of each type of application in the H-type application.
  • the threshold of each type of application in the H-type application is sent to the N mobile devices, so that the second mobile device monitors the total background power consumption of the first application.
  • the second mobile device may freeze, kill, or restart the first application.
  • the first application is any application that belongs to the H-type application on the second mobile device
  • the second threshold is a threshold of the application class to which the first application belongs.
  • the network device may periodically acquire the third data, where the third data includes data of the K mobile devices, which may be actively acquired by the network device, or may be actively sent by the K mobile devices.
  • the data of the third mobile device includes the duration of the application of the device on the third mobile device, and the period of the data collected by the third mobile device may be one hour, two hours, one day, and the like.
  • the duration of application of the device on the third mobile device may be the duration of use of all devices by all applications on the third mobile device, or the duration of use of the device by the third mobile device using the highest power consumption of all applications; It can be the duration that the third mobile device uses the B devices in all devices using the most power-intensive A applications of all applications.
  • the third data is the duration of 12 of the 10 most power-hungry applications (DISPLAY, CPU, GNSS, SENSOR, GPU, FRONTCAM, REARCAM, FLASHLIGHT, AUDIO, MODEM, WIFI, BT).
  • the duration of the device is the total length of use of the front and back of the device, and the third mobile device is any of the K mobile devices, and K is an integer greater than 2.
  • the threshold value applied to the K mobile devices is determined according to the third data, and the second feature vector may be determined according to the third data, where the second feature vector is determined in detail.
  • the process is the same as the process of determining the first feature vector in step 102. For detailed description, refer to step 102, and details are not described herein again.
  • the thresholds applied on the K mobile devices are then determined according to the second feature vector and the SVM, or the thresholds applied on the K mobile devices are determined according to the second feature vector and the DNN model.
  • the process of determining the third threshold according to the second feature vector and the SVM, or determining the third threshold according to the second feature vector and the DNN model is the same as the process of determining the predicted power consumption according to the first feature vector and the linear regression model in step 102,
  • the process of determining the predicted power consumption according to the first feature vector and the linear regression model in step 102 please refer to step 102, and details are not described herein again.
  • the network device after determining the threshold applied by the K mobile devices according to the third data, sends the thresholds applied by the K mobile devices to the K mobile devices, so that the third mobile device monitors the power consumption of the third application.
  • the third mobile device may freeze, kill, or restart the third application.
  • the third application is an application on the third mobile device, and the third threshold is a threshold of the third application.
  • steps 701-704, 705-708, and steps 709-711 may be performed in parallel or serially. Steps 705-708 and steps 709-711 may exist, and none of them may exist, or only one of them may exist.
  • FIG. 13 is a schematic structural diagram of a network device according to an embodiment of the present invention.
  • the network device may include:
  • the communication unit 1301 is configured to acquire first data, where the first data includes data of the M mobile devices, where the data of the first mobile device includes power consumption and duration of the first mobile device using the device, and power consumption of the first mobile device, The duration of the bright screen and the duration of the screen is off.
  • the first mobile device is any one of the M mobile devices, and M is an integer greater than 2.
  • the first determining unit 1302 is configured to determine, according to the first data acquired by the communication unit 1301, the predicted power consumption;
  • a second determining unit 1303, configured to determine a first threshold according to actual power consumption of the M mobile devices and predicted power consumption determined by the first determining unit 1302;
  • the communication unit 1301 is further configured to send the predicted power consumption determined by the first determining unit 1302 and the first threshold determined by the second determining unit 1303 to the M mobile devices, and the predicted power consumption and the first threshold are used to indicate the first mobile device.
  • the actual power consumption of the first mobile device is monitored. When the difference between the actual power consumption of the first mobile device and the predicted power consumption is greater than the first threshold, determining that the total power consumption of the first mobile device is in an abnormal state.
  • FIG. 14 is a schematic structural diagram of another network device according to an embodiment of the present invention.
  • the network device shown in FIG. 14 is optimized by the network device shown in FIG.
  • the first determining unit 1302 includes:
  • the predicted power consumption is determined based on the first eigenvector and the linear regression model.
  • the second determining unit 1303 may include:
  • a calculating subunit 13031 configured to calculate a difference between an actual power consumption and a predicted power consumption of each of the M mobile devices to obtain a power consumption difference value
  • the normal subunit 13032 is configured to perform normal distribution processing on the power consumption difference calculated by the calculation subunit 13031 to obtain a normal distribution difference value;
  • the first determining sub-unit 13033 is configured to determine a first threshold from the normal distribution difference values obtained from the normal sub-unit 13032 according to a preset rule.
  • the communication unit 1301 is further configured to acquire second data, where the second data includes data of the N mobile devices, and the data of the second mobile device includes a background total work of the application using the device on the second mobile device.
  • the second mobile device is any one of N mobile devices, and N is an integer greater than 2;
  • the network device can also include:
  • a classifying unit 1304 configured to classify applications on the N mobile devices according to the second data acquired by the communication unit 1301 to obtain a class H application, where H is an integer greater than two;
  • a third determining unit 1305, configured to determine a threshold of each type of application in the class H application classified by the classifying unit 1304;
  • the communication unit 1301 is further configured to send, by the third determining unit 1305, a threshold of each type of application in the H-type application to the N mobile devices, where the second threshold is used to instruct the second mobile device to monitor the total background power consumption of the first application.
  • the second threshold is used to instruct the second mobile device to monitor the total background power consumption of the first application.
  • the classification unit 1304 may include:
  • a drawing subunit 13041 configured to draw a CDF curve of total background power consumption applied on the N mobile devices according to the second data
  • a second determining subunit 13042 configured to determine, according to the CDF curve drawn by the drawing subunit 13041, the number of categories H applied on the N mobile devices;
  • the classification sub-unit 13043 is configured to divide the applications on the N mobile devices into the H-class determined by the second determining sub-unit 13042 according to the CDF curve and the CNN model drawn by the rendering sub-unit 13041.
  • the third determining unit 1305 includes:
  • the background total power consumption of the second application on the L mobile devices in the N mobile devices is determined as the threshold of the target class application, and the total background power consumption of the second application on the L mobile devices is the same, L and N
  • the ratio is equal to the preset ratio
  • the target application is any type of application in the H-type application
  • the second application is the application with the highest power consumption in the target application.
  • the communication unit 1301 is further configured to acquire third data, where the third data includes data of the K mobile devices, and the data of the third mobile device includes a duration of the application device used by the third mobile device, where The third mobile device is any one of the K mobile devices, and K is an integer greater than 2;
  • the network device can also include:
  • a fourth determining unit 1306, configured to determine, according to the third data acquired by the communication unit 1301, a threshold applied on the K mobile devices;
  • the communication unit 1301 is further configured to send, by the fourth determining unit 1306, a threshold value applied to the K mobile devices to the K mobile devices, where the third threshold is used to instruct the third mobile device to monitor the power consumption of the third application, where When the difference between the power consumption of the third application and the third threshold is greater than a preset value, determining that the power consumption of the third application is in an abnormal state, the third application is an application on the third mobile device, and the third threshold is a threshold of the third application. .
  • the fourth determining unit 1306 includes:
  • the thresholds applied on the K mobile devices are determined according to the second feature vector and the branch SVM, or the thresholds applied on the K mobile devices are determined according to the second feature vector and the DNN model.
  • FIG. 15 is a schematic structural diagram of still another network device according to an embodiment of the present invention.
  • the network device can include a processor 1501, a memory 1502, a transceiver 1503, and a bus 1504.
  • the processor 1501 may be a general purpose central processing unit (CPU), a plurality of CPUs, a microprocessor, an application-specific integrated circuit (ASIC), or one or more for controlling the execution of the program of the present invention. integrated circuit.
  • the memory 1002 may be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type that can store information and instructions.
  • ROM read-only memory
  • RAM random access memory
  • the dynamic storage device can also be an Electrically Erasable Programmable Read-Only Memory (EEPROM), a Compact Disc Read-Only Memory (CD-ROM) or other optical disc storage, and a disc storage device. (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or can be used to carry or store desired program code in the form of instructions or data structures and can be Any other media accessed, but not limited to this.
  • the memory 1502 may exist independently, and the bus 1504 is connected to the processor 1501.
  • the memory 1502 can also be integrated with the processor 1501.
  • Bus 1504 can include a path for communicating information between the components described above. among them:
  • the transceiver 1503 is configured to acquire first data, where the first data includes data of the M mobile devices, where the data of the first mobile device includes power consumption and duration of the first mobile device using the device, and power consumption of the first mobile device, The duration of the bright screen and the duration of the screen is off.
  • the first mobile device is any one of the M mobile devices, and M is an integer greater than 2.
  • the memory 1502 stores a set of program codes, and the processor 1501 is configured to call the program code stored in the memory 1502 to perform the following operations:
  • the transceiver 1503 is further configured to send the predicted power consumption and the first threshold to the M mobile devices, and the predicted power consumption and the first threshold are used to instruct the first mobile device to monitor the actual power consumption of the first mobile device, when the first mobile When the difference between the actual power consumption of the device and the predicted power consumption is greater than the first threshold, it is determined that the total power consumption of the first mobile device is in an abnormal state.
  • the determining, by the processor 1501, the predicted power consumption according to the first data includes:
  • the predicted power consumption is determined based on the first eigenvector and the linear regression model.
  • determining, by the processor 1501, the first threshold according to actual power consumption and predicted power consumption of the M mobile devices includes:
  • the first threshold is determined from the normal distribution difference according to a preset rule.
  • the transceiver 1503 is further configured to acquire second data, where the second data includes data of the N mobile devices, and the data of the second mobile device includes a background total work of the application used by the second mobile device. Consumption, the second mobile device is any one of N mobile devices, and N is an integer greater than 2;
  • the processor 1501 is further configured to call the program code stored in the memory 1502 to perform the following operations:
  • the transceiver 1503 is further configured to send a threshold of each type of application in the H-type application to the N mobile devices, where the second threshold is used to instruct the second mobile device to monitor the total background power consumption of the first application, when the first application is in the background.
  • the second threshold is used to instruct the second mobile device to monitor the total background power consumption of the first application, when the first application is in the background.
  • the processor 1501 classifies the applications on the N mobile devices according to the second data, to obtain the H-type applications, including:
  • the applications on the N mobile devices are classified into H categories according to the CDF curve and the CNN model.
  • the processor 1501 determines thresholds for each type of application in the H-type application, including:
  • the background total power consumption of the second application on the L mobile devices in the N mobile devices is determined as the threshold of the target class application, and the total background power consumption of the second application on the L mobile devices is the same, L and N
  • the ratio is equal to the preset ratio
  • the target application is any type of application in the H-type application
  • the second application is the application with the highest power consumption in the target application.
  • the transceiver 1503 is further configured to acquire third data, where the third data includes data of the K mobile devices, and the data of the third mobile device includes a duration of the device used by the third mobile device, where The third mobile device is any one of the K mobile devices, and K is an integer greater than 2;
  • the processor 1501 is further configured to call the program code stored in the memory 1502 to perform the following operations:
  • the transceiver 1503 is further configured to send a threshold of the application on the K mobile devices to the K mobile devices, where the third threshold is used to instruct the third mobile device to monitor the power consumption of the third application, when the power consumption of the third application is When the difference between the three thresholds is greater than the preset value, it is determined that the power consumption of the third application is in an abnormal state, the third application is an application on the third mobile device, and the third threshold is a threshold of the third application.
  • the determining, by the processor 1501, the thresholds applied on the K mobile devices according to the third data includes:
  • a threshold applied on the K mobile devices is determined according to the second feature vector and the SVM, or a threshold applied on the K mobile devices is determined according to the second feature vector and the DNN model.
  • steps 102-103, steps 702-703, steps 706-707, and step 710 may be performed by the processor 1501 and the memory 1502 in the network device, step 101, step 104, step 701, steps 704-705, step 708. -709 and step 711 can be performed by transceiver 1503 in the network device.
  • the first determining unit 1302, the second determining unit 1303, the classifying unit 1304, the third determining unit 1305, and the fourth determining unit 1306 may be implemented by a processor 1501 and a memory 1502 in the network device, and the communication unit 1301 may be configured by the network.
  • the transceiver 1503 in the device is implemented.
  • the embodiment of the invention also discloses a readable storage medium storing program code for executing the data processing method shown in Figs. 1 and 7 by the network device.
  • the functions described herein can be implemented in hardware, software, firmware, or any combination thereof.
  • the functions may be stored in a computer readable medium or transmitted as one or more instructions or code on a computer readable medium.
  • Computer readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one location to another.
  • a storage medium may be any available media that can be accessed by a general purpose or special purpose computer.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Telephone Function (AREA)

Abstract

一种数据处理方法，该方法应用于网络设备，获取包括M个移动设备的数据的第一数据，第一移动设备的数据包括第一移动设备使用器件的功耗和时长，以及第一移动设备的功耗、亮屏时长和灭屏时长，第一移动设备是是M个移动设备中的任一移动设备，M为大于2的整数；根据第一数据确定预测功耗；根据M个移动设备的实际功耗和预测功耗确定第一阈值；将预测功耗和第一阈值发送给M个移动设备，预测功耗和第一阈值用于指示第一移动设备监测第一移动设备的实际功耗，当第一移动设备的实际功耗与预测功耗的差值大于第一阈值时，确定第一移动设备的总功耗处于异常状态。本发明实施例，可以根据第一阈值和预测功耗对M个移动设备的功耗进行系统有效的管控。

Description

一种数据处理方法及网络设备 技术领域
本发明实施例涉及数据处理领域，具体涉及一种数据处理方法及网络设备。
背景技术
Android系统由于其开放性和多任务性给用户带来了许多便宜，如使用方便、成本低廉等，但同时也带来了诸多问题，如第三方应用的功耗问题。由于第三方应用各行其道，以致Android系统的功耗较大。目前，Android系统主要通过后台top应用监测第三方应用的功耗，但由于没有对移动设备功耗进行全面监测的策略，以致无法对移动设备功耗进行系统有效的管控。
发明内容
本发明实施例公开了一种数据处理方法及网络设备，用于提供一种移动设备功耗全面监测策略，以便可以对移动设备的功耗进行系统有效的管控。
第一方面公开数据处理方法，该方法应用于网络设备，获取包括M个移动设备的数据的第一数据，根据第一数据确定预测功耗，根据M个移动设备的实际功耗和预测功耗确定第一阈值，将预测功耗和第一阈值发送给M个移动设备，预测功耗和第一阈值用于指示第一移动设备监测第一移动设备的实际功耗，当第一移动设备的实际功耗与预测功耗的差值大于第一阈值时，确定第一移动设备的总功耗处于异常状态，以便可以根据第一阈值和预测功耗对这M个移动设备的功耗进行系统有效的管控。其中，第一移动设备的数据包括第一移动设备使用器件的功耗和时长，以及第一移动设备的功耗、亮屏时长和灭屏时长，第一移动设备是M个移动设备中的任一移动设备，M为大于2的整数。
在一个实施例中，根据第一数据确定预测功耗时，可以先根据第一数据确定第一特征向量，之后根据第一特征向量和线性回归模型确定预测功耗，以便可以快速地确定预测功耗。
在一个实施例中，根据M个移动设备的实际功耗和预测功耗确定第一阈值时，可以先计算M个移动设备中的每个移动设备的实际功耗和预测功耗的差值以获得功耗差值，之后将功耗差值进行正态分布处理以获得正态分布差值，根据预设规则从正态分布差值中确定第一阈值，以便可以得到精度较高的第一阈值。
在一个实施例中，获取包括N个移动设备的数据的第二数据，根据第二数据对N个移动设备上的应用进行分类以获得H类应用，确定H类应用中每类应用的阈值，将H类应用中每类应用的阈值发送给N个移动设备，第二阈值用于指示第二移动设备监测第一应用的后台总功耗，当第一应用的后台总功耗大于第二阈值时，确定第一应用的功耗处于异常状态，以便可以根据H类应用中每类应用的阈值对不同类的应用进行系统有效的管控，同时可以动态调整每类应用的阈值。其中，第二移动设备的数据包括第二移动设备上应用使用器件的后台总功耗，第二移动设备是N个移动设备中的任一移动设备，N为大于2的整数，H为大于2的整数，第一应用是第二移动设备上属于H类应用中的任一应用，第二阈值是第一应用 所属应用类的阈值。
在一个实施例中，根据第二数据对N个移动设备上的应用进行分类以获得H类应用时，可以先根据第二数据绘制N个移动设备上应用的后台总功耗的累计分布函数(Cumulative Distribution Function，CDF)曲线，根据CDF曲线确定N个移动设备上应用的类别数H，根据CDF曲线和卷积神经网络(Convolutional Neural Network，CNN)将N个移动设备上的应用分为H类。
在一个实施例中，确定H类应用中每类应用的阈值，即将N个移动设备中的L个移动设备上的第二应用的后台总功耗确定为目标类应用的阈值，L个移动设备上的第二应用的后台总功耗均相同，L与N的比值等于预设比例，目标类应用是H类应用中的任一类应用，第二应用是目标类应用中功耗最大的应用。
在一个实施例中，获取包括K个移动设备的数据的第三数据，根据第三数据确定K个移动设备上应用的阈值，将K个移动设备上应用的阈值发送给K个移动设备，第三阈值用于指示第三移动设备监测第三应用的功耗，当第三应用的功耗与第三阈值的差值大于预设值时，确定第三应用的功耗处于异常状态，以便根据K个移动设备上应用的阈值对每个应用进行系统有效的管控。其中，第三移动设备的数据包括第三移动设备上应用使用器件的时长，第三移动设备是K个移动设备中的任一移动设备，K为大于2的整数，第三应用是第三移动设备上的应用，第三阈值是第三应用的阈值。
在一个实施例中，根据第三数据确定K个移动设备上应用的阈值时，可以先根据第三数据确定第二特征向量，之后根据第二特征向量和支持向量机(Support Vector Machine，SVM)确定K个移动设备上应用的阈值，或根据第二特征向量和深度神经网络(Deep Neural Network，DNN)模型确定K个移动设备上应用的阈值。
第二方面公开一种网络设备，该网络设备包括用于执行第一方面或第一方面的任一种可能实现方式所提供的数据处理方法的单元。
第三方面公开一种网络设备，该网络设备包括处理器、存储器和收发器，存储器用于存储程序代码，处理器用于执行程序代码，收发器用于与移动设备进行通信。当处理器执行存储器存储的程序代码时，使得处理器执行第一方面或第一方面的任一种可能实现方式所公开的数据处理方法。
第四方面公开一种可读存储介质，该可读存储介质存储了网络设备用于执行第一方面或第一方面的任一种可能实现方式所公开的数据处理方法的程序代码。
附图说明
图1是本发明实施例公开的一种数据处理方法的流程示意图；
图2是本发明实施例公开的一种功耗差值的直方图；
图3是本发明实施例公开的一种功耗差值的正态分布图；
图4是本发明实施例公开的一种箱型图原理的示意图；
图5是本发明实施例公开的一种箱型图确定的异常值的示意图；
图6是本发明实施例公开的一种耗电量随时间变化的示意图；
图7是本发明实施例公开的另一种数据处理方法的流程示意图；
图8是本发明实施例公开的一种CDF曲线的示意图；
图9是本发明实施例公开的一种CDF面积的直方图；
图10是本发明实施例公开的一种CDF面积的正态分布图；
图11是本发明实施例公开的一种轮廓系数的示意图；
图12是本发明实施例公开的一种分类后的应用的CDF曲线示意图；
图13是本发明实施例公开的一种网络设备的结构示意图；
图14是本发明实施例公开的另一种网络设备的结构示意图；
图15是本发明实施例公开的又一种网络设备的结构示意图。
具体实施方式
本发明实施例公开了一种数据处理方法及网络设备，用于提供一种移动设备功耗全面监测策略，以便可以对移动设备的功耗进行系统有效的管控。以下进行详细说明。
请参阅图1，图1是本发明实施例公开的一种数据处理方法的流程示意图。其中，该数据处理方法是从网络设备的角度来描述的。如图1所示，该数据处理方法可以包括以下步骤。
101、获取第一数据。
本实施例中，网络设备可以周期性地获取第一数据，第一数据包括M个移动设备的数据，可以是网络设备主动获取的，也可以是M个移动设备主动发送的。其中，第一移动设备的数据包括第一移动设备使用器件的功耗和时长，以及第一移动设备的功耗、亮屏时长和灭屏时长，第一移动设备采集数据的周期可以为一小时、两小时、一天等。第一移动设备使用器件的功耗和时长可以是第一移动设备上所有应用使用所有器件的功耗和时长；也可以是第一移动设备使用所有器件的功耗和时长；也可以是第一移动设备上所有应用中功耗最大的A个应用使用器件的功耗和时长，A为大于或等于2的整数；也可以是第一移动设备使用所有器件中功耗最大的B个器件的功耗和时长，B为大于或等于2的整数；也可以是第一移动设备上所有应用中功耗最大的A个应用使用所有器件中的B个器件的功耗和时长。例如：第一数据为每天使用的最耗电的10个应用中的12个器件(DISPLAY、CPU、GNSS、SENSOR、GPU、FRONTCAM、REARCAM、FLASHLIGHT、AUDIO、MODEM、WIFI、BT)的功耗和时长。其中，器件的功耗和时长是器件的前后台总功耗和前后台使用总时长。其中，第一移动设备是与网络设备建立连接的M个移动设备中的任一移动设备，M为大于2的整数。
本实施例中，网络设备获取到第一数据之后，可以先对第一数据进行优化处理，即过滤掉第一数据中的异常数据，也即过滤掉第一数据中的噪声。当M个移动设备采集数据的周期为一天时，对第一数据进行优化处理包括：统计时长要等于充电时长与非充电时长之和，或等于灭屏时长与亮屏时长之和；统计时长大于或等于22小时且小于或等于26小时；第一移动设备每天的功耗为亮屏功耗与灭屏功耗之和除以电池容量再乘以100；充电时长与统计时长的比例要小于或等于0.3333；第一移动设备每天的功耗要大于或等于5格电量，其中，第一移动设备的总电量分为100格。
102、根据第一数据确定预测功耗。
本实施例中，网络设备获取到第一数据之后，将根据第一数据确定预测功耗，即根据第一数据确定第一特征向量，以及根据第一特征向量和线性回归模型确定预测功耗。其中，此处的第一数据可以是网络设备直接获取的第一数据，也可以是经过优化处理后的第一数据。
本实施例中，网络设备获取到第一数据之后，可以先确定特征向量的维度。举例说明，当M个移动设备的系统为126版本，且M个移动设备采集数据的周期为一天时，假设M个移动设备每天使用器件的功耗和时长为每个移动设备上每天使用所有应用中功耗最大的100个应用使用所有器件中的12个器件的功耗和时长，12个器件分别为DISPLAY、CPU、GNSS、SENSOR、GPU、FRONTCAM、REARCAM、FLASHLIGHT、AUDIO、MODEM、WIFI、BT。此时，向量维度为2830(100*14*2+14*2+2)，其中，100*14*2中的100为应用数量，14为器件数量，由于MODEM和WIFI分为上下行，因此，器件数量不是12，而是14，2表示采集的是前后台的数据；14*2用于表征除上述100应用之外其他应用所使用这十二个器件的数据维度，其中，14和2的含义与之前的相同；最后的2表示采集的是亮屏和灭屏的时长。
本实施例中，网络设备确定特征向量的维度之后，将根据第一数据和特征向量的维度确定第三特征向量。其中，对未使用的器件的数据进行补零处理，以便确保每个向量的维度相同。之后对第三特征向量进行标准化或归一化处理得到第四特征向量，以便对第三特征向量中的数据去量纲，从而将第三特征向量中有单位的数据统一为无单位的数据。之后去掉第四特征向量中相关性较小的特征得到第一特征向量，即当特征相关性较小时，将删除该特征。之后根据第一特征向量和线性回归模型确定预测功耗，即将第一特征向量作为线性回归模型的输入确定预测功耗。在确定预测功耗时，可以尝试不同参数下的结果，选择误差最小的参数确定的线性回归模型的结果为预测功耗。
103、根据M个移动设备的实际功耗和预测功耗确定第一阈值。
本实施例中，网络设备根据第一数据确定预测功耗之后，将根据M个移动设备的实际功耗和预测功耗确定第一阈值。可以先计算M个移动设备中的每个移动设备的实际功耗和预测功耗的差值以获得功耗差值，之后将功耗差值进行正态分布处理以获得正态分布差值。请参阅图2，图2是本发明实施例公开的一种功耗差值的直方图，如图2所示，横坐标为功耗差值dv，纵坐标为功耗差值出现的频率，可见，功耗差值的直方图的偏度小于1，但峰值远大于1，不能近似为正态分布。请参阅图3，图3是本发明实施例公开的一种功耗差值的正态分布图。如图3所示，图3是由图2所示的功耗差值的直方图经过正态分布处理后的正态分布图。之后根据预设规则从正态分布差值中确定第一阈值，可以根据统计学中的3σ准则，将图3中处于μ+3σ处的差值确定为第一阈值。
本实施例中，也可以采用箱型图来确定第一阈值，请参阅图4，图4是本发明实施例公开的一种箱型图原理的示意图。如图4所示：下四分位数为Q1，中位数为Q2，上四分位数为Q3，IRQ＝Q3-Q1，将功耗差值处于[Q1-3IQR，Q3+3IQR]区间之外的值确定为重度异常值，将功耗差值处于[Q1-1.5IQR，Q3+1.5IQR]区间之外的值确定为轻度异常值。请参阅图5，图5是本发明实施例公开的一种箱型图确定的异常值的示意图。如图5所示，纵坐标为功耗差值，最下面的细横线为下限，中间三根横线中处于最下面的为下四分位数、处于中间的 为中位数、处于最上面的为上四分位数，最上面的细横线为上限。请参阅图6，图6是本发明实施例公开的一种功耗随时间变化的示意图。如图6所示，横坐标为时间(单位为秒)，纵坐标为功耗(单位为mAh)，可见，随时间的增加，移动设备的功耗越来越大，移动设备的总功耗的异常状态也越来越严重。
104、将预测功耗和第一阈值发送给M个移动设备。
本实施例中，网络设备根据M个移动设备的实际功耗和预测功耗确定第一阈值之后，将预测功耗和第一阈值分别发送给M个移动设备，以便第一移动设备监测第一移动设备的实际功耗，当第一移动设备的实际功耗与预测功耗的差值大于第一阈值时，确定第一移动设备的总功耗处于异常状态，第一移动设备可以冷冻掉、杀掉或重启第一移动设备上的部分或全部后台应用。
请参阅图7，图7是本发明实施例公开的另一种数据处理方法的流程示意图。其中，该数据处理方法是从网络设备的角度来描述的。如图7所示，该数据处理方法可以包括以下步骤。
701、获取第一数据。
其中，步骤701与步骤101相同，详细描述请参考步骤101，在此不再赘述。
702、根据第一数据确定预测功耗。
其中，步骤702与步骤102相同，详细描述请参考步骤102，在此不再赘述。
703、根据M个移动设备的实际功耗和预测功耗确定第一阈值。
其中，步骤703与步骤103相同，详细描述请参考步骤103，在此不再赘述。
704、将预测功耗和第一阈值发送给M个移动设备。
其中，步骤704与步骤104相同，详细描述请参考步骤104，在此不再赘述。
705、获取第二数据。
本实施例中，网络设备可以周期性地获取第二数据，第二数据包括N个移动设备的数据，可以是网络设备主动获取的，也可以是N个移动设备主动发送的。其中，第二移动设备的数据包括第二移动设备上应用使用器件的后台总功耗，第二移动设备采集数据的周期可以为一小时、两小时、一天等。第二移动设备上应用使用器件的后台总功耗，可以是第二移动设备上所有应用使用所有器件的后台总功耗，也可以是第二移动设备上使用所有应用中功耗最大的A个应用使用器件的后台总功耗；也可以是第二移动设备上使用所有应用中功耗最大的A个应用使用所有器件中的B个器件的后台总功耗。例如：第二数据为每天使用的最耗电的10个应用中的12个器件(DISPLAY、CPU、GNSS、SENSOR、GPU、FRONTCAM、REARCAM、FLASHLIGHT、AUDIO、MODEM、WIFI、BT)的后台总功耗。其中，第二移动设备是N个移动设备中的任一移动设备，N为大于2的整数。
706、根据第二数据对N个移动设备上的应用进行分类，以获得H类应用。
本实施例中，网络设备获取到第二数据之后，将根据第二数据将N个移动设备上的应用分为H类，可以先根据第二数据绘制N个移动设备上应用的后台总功耗的CDF曲线。请参阅图8，图8是本发明实施例公开的一种CDF曲线的示意图。如图8所示，CDF曲线的横坐标为功耗(单位为mAh)，纵坐标为累计百分比(单位为％)。之后根据CDF曲线确定N个移动设备上应用的类别数H，可以先计算图8中第一区域的面积得到CDF面积。请参阅图9，图9 是本发明实施例公开的一种CDF面积的直方图。如图9所示，横坐标为CDF面积area，纵坐标为CDF面积出现的频率。请参阅图10，图10是本发明实施例公开的一种CDF面积的正态分布图。图10是由图9中CDF面积的直方图经过正态分布处理后得到的。之后对图10所示的CDF面积进行kmean聚类得到轮廓系数。请参阅图11，图11是本发明实施例公开的一种轮廓系数的示意图。如图11所示，横坐标为聚类的类别数，纵坐标为轮廓系数，从图11所示的曲线中选取轮廓系数的局部峰值对应的类别数分别为7、8、11、17和20，之后按照这几种类别数和CDF面积分别对应用进行分类。举例说明，假设应用分类的类别数为7，CDF面积为10-101，用最大CDF面积101减去最小CDF面积得到91，再用91除以7得到13，可以分别将CDF面积处于10-23之间、23-36之间、36-49之间、49-62之间、62-75之间、75-88之间和88-101之间的应用确定为同一类应用。可以将上述几种分类中应用的分布最均匀的类别数确定为最优的类别数。之后根据CDF曲线和CNN模型将N个移动设备上的应用分为H类，即对分类后的应用的CDF曲线打标签，将属于同一类的应用曲线标记出来，同时将不同类应用的CDF曲线区别开来。请参阅图12，图12是本发明实施例公开的一种分类后的应用的CDF曲线示意图。如图12所示，应用分类的类别数为10，每类应用均是CDF曲线相隔最近的应用。其中，H为大于2的整数。
707、确定H类应用中每类应用的阈值。
本实施例中，网络设备根据第二数据对N个移动设备上的应用进行分类得到H类应用之后，将确定H类应用中每类应用的阈值，可以将N个移动设备中的L个移动设备上的第二应用的后台总功耗确定为目标类应用的阈值，L个移动设备上的第二应用的后台总功耗均相同，L与N的比值等于预设比例，目标类应用是H类应用中的任一类应用，第二应用是目标类应用中功耗最大的应用。如图12所示，图中的横线为纵坐标80％对应的横线，可以将该横线与CDF曲线相交对应的横坐标的值确定为H类应用中每类应用的阈值。由于横线与H类应用中每类应用的交点不止一个，可以将每类应用中交点对应的最大的横坐标值确定为H类应用中每类应用的阈值。
708、将H类应用中每类应用的阈值发送给N个移动设备。
本实施例中，网络设备确定H类应用中每类应用的阈值之后，将H类应用中每类应用的阈值发送给N个移动设备，以便第二移动设备监测第一应用的后台总功耗，当第一应用的后台总功耗大于第二阈值时，确定第一应用的功耗处于异常状态，第二移动设备可以冷冻掉、杀掉或重启第一应用。其中，第一应用是第二移动设备上属于H类应用中的任一应用，第二阈值是第一应用所属应用类的阈值。
709、获取第三数据。
本实施例中，网络设备可以周期性地获取第三数据，第三数据包括K个移动设备的数据，可以是网络设备主动获取的，也可以是K个移动设备主动发送的。其中，第三移动设备的数据包括第三移动设备上应用使用器件的时长，第三移动设备采集数据的周期可以为一小时、两小时、一天等。第三移动设备上应用使用器件的时长，可以是第三移动设备上所有应用使用所有器件的时长，也可以是第三移动设备使用所有应用中功耗最大的A个应用使用器件的时长；也可以是第三移动设备使用所有应用中功耗最大的A个应用使用所有器件中的B个器件的时长。例如：第三数据为每天使用的最耗电的10个应用中的12个器件 (DISPLAY、CPU、GNSS、SENSOR、GPU、FRONTCAM、REARCAM、FLASHLIGHT、AUDIO、MODEM、WIFI、BT)的时长。其中，器件的时长是器件的前后台使用总时长，第三移动设备是K个移动设备中的任一移动设备，K为大于2的整数。
710、根据第三数据确定K个移动设备上应用的阈值。
本实施例中，网络设备获取到第三数据之后，将根据第三数据确定K个移动设备上应用的阈值，可以先根据第三数据确定第二特征向量，其中，确定第二特征向量的详细过程与步骤102中确定第一特征向量的过程相同，详细描述请参考步骤102，在此不再赘述。之后根据第二特征向量和SVM确定K个移动设备上应用的阈值，或根据第二特征向量和DNN模型确定K个移动设备上应用的阈值。其中，根据第二特征向量和SVM确定第三阈值，或根据第二特征向量和DNN模型确定第三阈值的过程与步骤102中根据第一特征向量和线性回归模型确定预测功耗的过程相同，详细描述请参考步骤102，在此不再赘述。
711、将K个移动设备上应用的阈值发送给K个移动设备。
本实施例中，网络设备根据第三数据确定K个移动设备上应用的阈值之后，将K个移动设备上应用的阈值发送给K个移动设备，以便第三移动设备监测第三应用的功耗，当第三应用的功耗与第三阈值的差值大于预设值时，确定第三应用的功耗处于异常状态，第三移动设备可以冷冻掉、杀掉或重启第三应用。其中，第三应用是第三移动设备上的应用，第三阈值是第三应用的阈值。
其中，步骤701-704、步骤705-708以及步骤709-711之间可以是并行执行的，也可以是串行执行的。步骤705-708与步骤709-711可以均存在，可以均不存在，也可以只存在其一。
请参阅图13，图13是本发明实施例公开的一种网络设备的结构示意图。如图13所示，该网络设备可以包括：
通信单元1301，用于获取第一数据，第一数据包括M个移动设备的数据，第一移动设备的数据包括第一移动设备使用器件的功耗和时长，以及第一移动设备的功耗、亮屏时长和灭屏时长，第一移动设备是M个移动设备中的任一移动设备，M为大于2的整数；
第一确定单元1302，用于根据通信单元1301获取的第一数据确定预测功耗；
第二确定单元1303，用于根据M个移动设备的实际功耗和第一确定单元1302确定的预测功耗确定第一阈值；
通信单元1301，还用于将第一确定单元1302确定的预测功耗和第二确定单元1303确定的第一阈值发送给M个移动设备，预测功耗和第一阈值用于指示第一移动设备监测第一移动设备的实际功耗，当第一移动设备的实际功耗与预测功耗的差值大于第一阈值时，确定第一移动设备的总功耗处于异常状态。
请参阅图14，图14是本发明实施例公开的另一种网络设备的结构示意图。其中，图14所示的网络设备是由图13所示的网络设备优化得到的。其中，第一确定单元1302包括：
根据第一数据确定第一特征向量；
根据第一特征向量和线性回归模型确定预测功耗。
作为一种可能的实施方式，第二确定单元1303可以包括：
计算子单元13031，用于计算M个移动设备中的每个移动设备的实际功耗和预测功耗的 差值，以获得功耗差值；
正态子单元13032，用于将计算子单元13031计算得到的功耗差值进行正态分布处理，以获得正态分布差值；
第一确定子单元13033，用于根据预设规则从正态子单元13032得到的正态分布差值中确定第一阈值。
作为一种可能的实施方式，通信单元1301，还用于获取第二数据，第二数据包括N个移动设备的数据，第二移动设备的数据包括第二移动设备上应用使用器件的后台总功耗，第二移动设备是N个移动设备中的任一移动设备，N为大于2的整数；
该网络设备还可以包括：
分类单元1304，用于根据通信单元1301获取的第二数据对N个移动设备上的应用进行分类，以获得H类应用，H为大于2的整数；
第三确定单元1305，用于确定分类单元1304分类的H类应用中每类应用的阈值；
通信单元1301，还用于将第三确定单元1305确定的H类应用中每类应用的阈值发送给N个移动设备，第二阈值用于指示第二移动设备监测第一应用的后台总功耗，当第一应用的后台总功耗大于第二阈值时，确定第一应用的功耗处于异常状态，第一应用是第二移动设备上属于H类应用中的任一应用，第二阈值是第一应用所属应用类的阈值。
作为一种可能的实施方式，分类单元1304可以包括：
绘制子单元13041，用于根据第二数据绘制N个移动设备上应用的后台总功耗的CDF曲线；
第二确定子单元13042，用于根据绘制子单元13041绘制的CDF曲线确定N个移动设备上应用的类别数H；
分类子单元13043，用于根据绘制子单元13041绘制的CDF曲线和CNN模型将N个移动设备上的应用分为第二确定子单元13042确定的H类。
作为一种可能的实施方式，第三确定单元1305包括：
将N个移动设备中的L个移动设备上的第二应用的后台总功耗确定为目标类应用的阈值，L个移动设备上的第二应用的后台总功耗均相同，L与N的比值等于预设比例，目标类应用是H类应用中的任一类应用，第二应用是目标类应用中功耗最大的应用。
作为一种可能的实施方式，通信单元1301，还用于获取第三数据，第三数据包括K个移动设备的数据，第三移动设备的数据包括第三移动设备上应用使用器件的时长，第三移动设备是K个移动设备中的任一移动设备，K为大于2的整数；
该网络设备还可以包括：
第四确定单元1306，用于根据通信单元1301获取的第三数据确定K个移动设备上应用的阈值；
通信单元1301，还用于将第四确定单元1306确定的K个移动设备上应用的阈值发送给K个移动设备，第三阈值用于指示第三移动设备监测第三应用的功耗，当第三应用的功耗与第三阈值的差值大于预设值时，确定第三应用的功耗处于异常状态，第三应用是第三移动设备上的应用，第三阈值是第三应用的阈值。
作为一种可能的实施方式，第四确定单元1306包括：
根据第三数据确定第二特征向量；
根据第二特征向量和支SVM确定K个移动设备上应用的阈值，或根据第二特征向量和DNN模型确定K个移动设备上应用的阈值。
请参阅图15，图15是本发明实施例公开的又一种网络设备的结构示意图。如图15所示，该网络设备可以包括处理器1501、存储器1502、收发器1503和总线1504。处理器1501可以是一个通用中央处理器(CPU)，多个CPU，微处理器，特定应用集成电路(application-specific integrated circuit，ASIC)，或一个或多个用于控制本发明方案程序执行的集成电路。存储器1002可以是只读存储器(read-only memory，ROM)或可存储静态信息和指令的其他类型的静态存储设备，随机存取存储器(random access memory，RAM)或者可存储信息和指令的其他类型的动态存储设备，也可以是电可擦可编程只读存储器(Electrically Erasable Programmable Read-Only Memory，EEPROM)、只读光盘(Compact Disc Read-Only Memory，CD-ROM)或其他光盘存储、光碟存储(包括压缩光碟、激光碟、光碟、数字通用光碟、蓝光光碟等)、磁盘存储介质或者其他磁存储设备、或者能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质，但不限于此。存储器1502可以是独立存在，总线1504与处理器1501相连接。存储器1502也可以和处理器1501集成在一起。总线1504可包括一通路，在上述组件之间传送信息。其中：
收发器1503，用于获取第一数据，第一数据包括M个移动设备的数据，第一移动设备的数据包括第一移动设备使用器件的功耗和时长，以及第一移动设备的功耗、亮屏时长和灭屏时长，第一移动设备是M个移动设备中的任一移动设备，M为大于2的整数；
存储器1502存储有一组程序代码，处理器1501用于调用存储器1502存储的程序代码执行以下操作：
根据第一数据确定预测功耗；
根据M个移动设备的实际功耗和预测功耗确定第一阈值；
收发器1503，还用于将预测功耗和第一阈值发送给M个移动设备，预测功耗和第一阈值用于指示第一移动设备监测第一移动设备的实际功耗，当第一移动设备的实际功耗与预测功耗的差值大于第一阈值时，确定第一移动设备的总功耗处于异常状态。
作为一种可能的实施方式，处理器1501根据第一数据确定预测功耗包括：
根据第一数据确定第一特征向量；
根据第一特征向量和线性回归模型确定预测功耗。
作为一种可能的实施方式，处理器1501根据M个移动设备的实际功耗和预测功耗确定第一阈值包括：
计算M个移动设备中的每个移动设备的实际功耗和预测功耗的差值，以获得功耗差值；
将功耗差值进行正态分布处理，以获得正态分布差值；
根据预设规则从正态分布差值中确定第一阈值。
作为一种可能的实施方式，收发器1503，还用于获取第二数据，第二数据包括N个移动设备的数据，第二移动设备的数据包括第二移动设备上应用使用器件的后台总功耗，第二移动设备是N个移动设备中的任一移动设备，N为大于2的整数；
处理器1501还用于调用存储器1502存储的程序代码执行以下操作：
根据第二数据对N个移动设备上的应用进行分类，以获得H类应用，H为大于2的整数；
确定H类应用中每类应用的阈值；
收发器1503，还用于将H类应用中每类应用的阈值发送给N个移动设备，第二阈值用于指示第二移动设备监测第一应用的后台总功耗，当第一应用的后台总功耗大于第二阈值时，确定第一应用的功耗处于异常状态，第一应用是第二移动设备上属于H类应用中的任一应用，第二阈值是第一应用所属应用类的阈值。
作为一种可能的实施方式，处理器1501根据第二数据对N个移动设备上的应用进行分类，以获得H类应用包括：
根据第二数据绘制N个移动设备上应用的后台总功耗的CDF曲线；
根据CDF曲线确定N个移动设备上应用的类别数H；
根据CDF曲线和CNN模型将N个移动设备上的应用分为H类。
作为一种可能的实施方式，处理器1501确定H类应用中每类应用的阈值包括：
将N个移动设备中的L个移动设备上的第二应用的后台总功耗确定为目标类应用的阈值，L个移动设备上的第二应用的后台总功耗均相同，L与N的比值等于预设比例，目标类应用是H类应用中的任一类应用，第二应用是目标类应用中功耗最大的应用。
作为一种可能的实施方式，收发器1503，还用于获取第三数据，第三数据包括K个移动设备的数据，第三移动设备的数据包括第三移动设备上应用使用器件的时长，第三移动设备是K个移动设备中的任一移动设备，K为大于2的整数；
处理器1501还用于调用存储器1502存储的程序代码执行以下操作：
根据第三数据确定K个移动设备上应用的阈值；
收发器1503，还用于将K个移动设备上应用的阈值发送给K个移动设备，第三阈值用于指示第三移动设备监测第三应用的功耗，当第三应用的功耗与第三阈值的差值大于预设值时，确定第三应用的功耗处于异常状态，第三应用是第三移动设备上的应用，第三阈值是第三应用的阈值。
作为一种可能的实施方式，处理器1501根据第三数据确定K个移动设备上应用的阈值包括：
根据第三数据确定第二特征向量；
根据第二特征向量和SVM确定K个移动设备上应用的阈值，或根据第二特征向量和DNN模型确定K个移动设备上应用的阈值。
其中，步骤102-103、步骤702-703、步骤706-707和步骤710可以由网络设备中的处理器1501和存储器1502来执行，步骤101、步骤104、步骤701、步骤704-705、步骤708-709和步骤711可以由网络设备中的收发器1503来执行。
其中，第一确定单元1302、第二确定单元1303、分类单元1304、第三确定单元1305和第四确定单元1306可以由网络设备中的处理器1501和存储器1502来实现，通信单元1301可以由网络设备中的收发器1503来实现。
本发明实施例还公开了一种可读存储介质，该可读存储介质存储了网络设备用于执行 图1和7所示的数据处理方法的程序代码。
本领域技术人员应该可以意识到，在上述一个或多个示例中，本发明所描述的功能可以用硬件、软件、固件或它们的任意组合来实现。当使用软件实现时，可以将这些功能存储在计算机可读介质中或者作为计算机可读介质上的一个或多个指令或代码进行传输。计算机可读介质包括计算机存储介质和通信介质，其中，通信介质包括便于从一个地方向另一个地方传送计算机程序的任何介质。存储介质可以是通用或专用计算机能够存取的任何可用介质。
以上所述的具体实施方式，对本发明的目的、技术方案和有益效果进行了进一步详细说明，所应理解的是，以上所述仅为本发明的具体实施方式而已，并不用于限定本发明的保护范围，凡在本发明的技术方案的基础之上，所做的任何修改、等同替换、改进等，均应包括在本发明的保护范围之内。

Claims (16)

  1. 一种数据处理方法，其特征在于，所述方法应用于网络设备，包括：
    获取第一数据，所述第一数据包括M个移动设备的数据，第一移动设备的数据包括所述第一移动设备使用器件的功耗和时长，以及所述第一移动设备的功耗、亮屏时长和灭屏时长，所述第一移动设备是所述M个移动设备中的任一移动设备，所述M为大于2的整数；
    根据所述第一数据确定预测功耗；
    根据所述M个移动设备的实际功耗和所述预测功耗确定第一阈值；
    将所述预测功耗和所述第一阈值发送给所述M个移动设备，所述预测功耗和所述第一阈值用于指示所述第一移动设备监测所述第一移动设备的实际功耗，当所述第一移动设备的实际功耗与所述预测功耗的差值大于所述第一阈值时，确定所述第一移动设备的总功耗处于异常状态。
  2. 根据权利要求1所述的方法，其特征在于，所述根据所述第一数据确定预测功耗包括：
    根据所述第一数据确定第一特征向量；
    根据所述第一特征向量和线性回归模型确定预测功耗。
  3. 根据权利要求1或2所述的方法，其特征在于，所述根据所述M个移动设备的实际功耗和所述预测功耗确定第一阈值包括：
    计算所述M个移动设备中的每个移动设备的实际功耗和所述预测功耗的差值，以获得功耗差值；
    将所述功耗差值进行正态分布处理，以获得正态分布差值；
    根据预设规则从所述正态分布差值中确定第一阈值。
  4. 根据权利要求1-3任一项所述的方法，其特征在于，所述方法还包括：
    获取第二数据，所述第二数据包括N个移动设备的数据，第二移动设备的数据包括所述第二移动设备上应用使用器件的后台总功耗，所述第二移动设备是所述N个移动设备中的任一移动设备，所述N为大于2的整数；
    根据所述第二数据对所述N个移动设备上的应用进行分类，以获得H类应用，所述H为大于2的整数；
    确定所述H类应用中每类应用的阈值；
    将所述H类应用中每类应用的阈值发送给所述N个移动设备，第二阈值用于指示所述第二移动设备监测第一应用的后台总功耗，当所述第一应用的后台总功耗大于所述第二阈值时，确定所述第一应用的功耗处于异常状态，所述第一应用是所述第二移动设备上属于所述H类应用中的任一应用，所述第二阈值是所述第一应用所属应用类的阈值。
  5. 根据权利要求4所述的方法，其特征在于，所述根据所述第二数据对所述N个移动设备上的应用进行分类，以获得H类应用包括：
    根据所述第二数据绘制所述N个移动设备上应用的后台总功耗的累计分布函数CDF曲线；
    根据所述CDF曲线确定所述N个移动设备上应用的类别数H；
    根据所述CDF曲线和卷积神经网络CNN模型将所述N个移动设备上的应用分为H类。
  6. 根据权利要求4或5所述的方法，其特征在于，所述确定所述H类应用中每类应用的阈值包括：
    将所述N个移动设备中的L个移动设备上的第二应用的后台总功耗确定为目标类应用的阈值，所述L个移动设备上的第二应用的后台总功耗均相同，所述L与所述N的比值等于预设比例，所述目标类应用是H类应用中的任一类应用，所述第二应用是所述目标类应用中功耗最大的应用。
  7. 根据权利要求1-6任一项所述的方法，其特征在于，所述方法还包括：
    获取第三数据，所述第三数据包括K个移动设备的数据，第三移动设备的数据包括所述第三移动设备上应用使用器件的时长，所述第三移动设备是所述K个移动设备中的任一移动设备，所述K为大于2的整数；
    根据所述第三数据确定所述K个移动设备上应用的阈值；
    将所述K个移动设备上应用的阈值发送给所述K个移动设备，第三阈值用于指示所述第三移动设备监测第三应用的功耗，当所述第三应用的功耗与所述第三阈值的差值大于预设值时，确定所述第三应用的功耗处于异常状态，所述第三应用是所述第三移动设备上的应用，所述第三阈值是所述第三应用的阈值。
  8. 根据权利要求7所述的方法，其特征在于，所述根据所述第三数据确定所述K个移动设备上应用的阈值包括：
    根据所述第三数据确定第二特征向量；
    根据所述第二特征向量和支持向量机SVM确定所述K个移动设备上应用的阈值，或根据所述第二特征向量和深度神经网络DNN模型确定所述K个移动设备上应用的阈值。
  9. 一种网络设备，其特征在于，包括处理器、存储器和收发器，其中：
    所述收发器，用于获取第一数据，所述第一数据包括M个移动设备的数据，第一移动设备的数据包括所述第一移动设备使用器件的功耗和时长，以及所述第一移动设备的功耗、亮屏时长和灭屏时长，所述第一移动设备是所述M个移动设备中的任一移动设备，所述M为大于2的整数；
    所述存储器存储有一组程序代码，所述处理器用于调用所述存储器存储的程序代码执行以下操作：
    根据所述第一数据确定预测功耗；
    根据所述M个移动设备的实际功耗和所述预测功耗确定第一阈值；
    所述收发器，还用于将所述预测功耗和所述第一阈值发送给所述M个移动设备，所述 预测功耗和所述第一阈值用于指示所述第一移动设备监测所述第一移动设备的实际功耗，当所述第一移动设备的实际功耗与所述预测功耗的差值大于所述第一阈值时，确定所述第一移动设备的总功耗处于异常状态。
  10. 根据权利要求9所述的网络设备，其特征在于，所述处理器根据所述第一数据确定预测功耗包括：
    根据所述第一数据确定第一特征向量；
    根据所述第一特征向量和线性回归模型确定预测功耗。
  11. 根据权利要求9或10所述的网络设备，其特征在于，所述处理器根据所述M个移动设备的实际功耗和所述预测功耗确定第一阈值包括：
    计算所述M个移动设备中的每个移动设备的实际功耗和所述预测功耗的差值，以获得功耗差值；
    将所述功耗差值进行正态分布处理，以获得正态分布差值；
    根据预设规则从所述正态分布差值中确定第一阈值。
  12. 根据权利要求9-11任一项所述的网络设备，其特征在于，所述收发器，还用于获取第二数据，所述第二数据包括N个移动设备的数据，第二移动设备的数据包括所述第二移动设备上应用使用器件的后台总功耗，所述第二移动设备是所述N个移动设备中的任一移动设备，所述N为大于2的整数；
    所述处理器还用于调用所述存储器存储的程序代码执行以下操作：
    根据所述第二数据对所述N个移动设备上的应用进行分类，以获得H类应用，所述H为大于2的整数；
    确定所述H类应用中每类应用的阈值；
    所述收发器，还用于将所述H类应用中每类应用的阈值发送给所述N个移动设备，第二阈值用于指示所述第二移动设备监测第一应用的后台总功耗，当所述第一应用的后台总功耗大于所述第二阈值时，确定所述第一应用的功耗处于异常状态，所述第一应用是所述第二移动设备上属于所述H类应用中的任一应用，所述第二阈值是所述第一应用所属应用类的阈值。
  13. 根据权利要求12所述的网络设备，其特征在于，所述处理器根据所述第二数据对所述N个移动设备上的应用进行分类，以获得H类应用包括：
    根据所述第二数据绘制所述N个移动设备上应用的后台总功耗的CDF曲线；
    根据所述CDF曲线确定所述N个移动设备上应用的类别数H；
    根据所述CDF曲线和CNN模型将所述N个移动设备上的应用分为H类。
  14. 根据权利要求12或13所述的网络设备，其特征在于，所述处理器确定所述H类应用中每类应用的阈值包括：
    将所述N个移动设备中的L个移动设备上的第二应用的后台总功耗确定为目标类应用的阈值，所述L个移动设备上的第二应用的后台总功耗均相同，所述L与所述N的比值等于预设比例，所述目标类应用是H类应用中的任一类应用，所述第二应用是所述目标类应用中功耗最大的应用。
  15. 根据权利要求9-14任一项所述的网络设备，其特征在于，所述收发器，还用于获取第三数据，所述第三数据包括K个移动设备的数据，第三移动设备的数据包括所述第三移动设备上应用使用器件的时长，所述第三移动设备是所述K个移动设备中的任一移动设备，所述K为大于2的整数；
    所述处理器还用于调用所述存储器存储的程序代码执行以下操作：
    根据所述第三数据确定所述K个移动设备上应用的阈值；
    所述收发器，还用于将所述K个移动设备上应用的阈值发送给所述K个移动设备，第三阈值用于指示所述第三移动设备监测第三应用的功耗，当所述第三应用的功耗与所述第三阈值的差值大于预设值时，确定所述第三应用的功耗处于异常状态，所述第三应用是所述第三移动设备上的应用，所述第三阈值是所述第三应用的阈值。
  16. 根据权利要求15所述的网络设备，其特征在于，所述处理器根据所述第三数据确定所述K个移动设备上应用的阈值包括：
    根据所述第三数据确定第二特征向量；
    根据所述第二特征向量和SVM确定所述K个移动设备上应用的阈值，或根据所述第二特征向量和DNN模型确定所述K个移动设备上应用的阈值。
PCT/CN2018/080568 2018-03-26 2018-03-26 一种数据处理方法及网络设备 WO2019183781A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2018/080568 WO2019183781A1 (zh) 2018-03-26 2018-03-26 一种数据处理方法及网络设备
CN201880090873.2A CN111819550B (zh) 2018-03-26 2018-03-26 一种数据处理方法及网络设备

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/080568 WO2019183781A1 (zh) 2018-03-26 2018-03-26 一种数据处理方法及网络设备

Publications (1)

Publication Number Publication Date
WO2019183781A1 true WO2019183781A1 (zh) 2019-10-03

Family

ID=68062446

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/080568 WO2019183781A1 (zh) 2018-03-26 2018-03-26 一种数据处理方法及网络设备

Country Status (2)

Country Link
CN (1) CN111819550B (zh)
WO (1) WO2019183781A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115002885A (zh) * 2022-06-28 2022-09-02 厦门汉印电子技术有限公司 标签数据采集方法、装置、电子设备及存储介质
CN116204566A (zh) * 2023-04-28 2023-06-02 深圳市欣冠精密技术有限公司 一种数字化工厂监控大数据处理系统

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113409578B (zh) * 2021-06-25 2022-05-13 北京航空航天大学 一种基于模糊聚类的交通网络健康画像方法及系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011053220A1 (en) * 2009-10-30 2011-05-05 Telefonaktiebolaget L M Ericsson (Publ) Method and apparatus for supporting uplink transmit diversity
CN105139038A (zh) * 2015-09-14 2015-12-09 李玮琛 肌电信号模式识别方法
CN105528374A (zh) * 2014-10-21 2016-04-27 苏宁云商集团股份有限公司 一种电子商务中的商品推荐方法及其系统
CN107817891A (zh) * 2017-11-13 2018-03-20 广东欧珀移动通信有限公司 屏幕控制方法、装置、设备及存储介质

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2010050249A1 (ja) * 2008-10-30 2010-05-06 株式会社日立製作所 情報処理システムの運用管理装置
CN101826996B (zh) * 2010-03-19 2012-05-23 中国科学院计算机网络信息中心 域名系统流量检测方法与域名服务器
US9098553B2 (en) * 2013-03-15 2015-08-04 Gridglo Llc System and method for remote activity detection
CN104166542B (zh) * 2013-05-17 2018-01-05 腾讯科技（深圳）有限公司 检测应用耗电的方法和装置
CN106383571A (zh) * 2016-09-09 2017-02-08 宇龙计算机通信科技(深圳)有限公司 一种应用程序处理装置及方法
CN106447198A (zh) * 2016-09-28 2017-02-22 国网山东省电力公司电力科学研究院 一种基于业扩报装数据的用电量核对方法
CN106814676B (zh) * 2017-03-14 2020-12-22 安徽德诺科技股份公司 楼宇能耗监控和管理的系统、方法和装置
CN107678803B (zh) * 2017-09-30 2020-12-01 Oppo广东移动通信有限公司 应用管控方法、装置、存储介质及电子设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011053220A1 (en) * 2009-10-30 2011-05-05 Telefonaktiebolaget L M Ericsson (Publ) Method and apparatus for supporting uplink transmit diversity
CN105528374A (zh) * 2014-10-21 2016-04-27 苏宁云商集团股份有限公司 一种电子商务中的商品推荐方法及其系统
CN105139038A (zh) * 2015-09-14 2015-12-09 李玮琛 肌电信号模式识别方法
CN107817891A (zh) * 2017-11-13 2018-03-20 广东欧珀移动通信有限公司 屏幕控制方法、装置、设备及存储介质

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115002885A (zh) * 2022-06-28 2022-09-02 厦门汉印电子技术有限公司 标签数据采集方法、装置、电子设备及存储介质
CN115002885B (zh) * 2022-06-28 2024-04-26 厦门汉印电子技术有限公司 标签数据采集方法、装置、电子设备及存储介质
CN116204566A (zh) * 2023-04-28 2023-06-02 深圳市欣冠精密技术有限公司 一种数字化工厂监控大数据处理系统

Also Published As

Publication number Publication date
CN111819550B (zh) 2022-04-05
CN111819550A (zh) 2020-10-23

Similar Documents

Publication Publication Date Title
WO2019183781A1 (zh) 一种数据处理方法及网络设备
CN108345524B (zh) 应用程序监控方法及应用程序监控装置
US9658910B2 (en) Systems and methods for spatially displaced correlation for detecting value ranges of transient correlation in machine data of enterprise systems
US7734768B2 (en) System and method for adaptively collecting performance and event information
CN107770088B (zh) 一种流量控制方法及装置
US7882160B2 (en) Low latency optimization for generational garbage collection
CN107608865B (zh) 数据存储方法和装置
CN104702680B (zh) 一种资源监测方法及系统
US10678596B2 (en) User behavior-based dynamic resource capacity adjustment
US11726540B2 (en) Usage scenario based monitoring and adjustment
CN109525500B (zh) 一种自调整阈值的信息处理方法及信息处理装置
TW201942739A (zh) 任務的分配方法、裝置及設備
CN110494848A (zh) 任务处理方法、设备及机器可读存储介质
CN114443429B (zh) 一种告警事件的处理方法、装置及计算机可读存储介质
CN105607952B (zh) 一种虚拟化资源的调度方法及装置
CN108809760A (zh) 数据采样系统中采样周期的控制方法和装置
CN111901405B (zh) 一种多节点监控方法、装置、电子设备及存储介质
CN109597800B (zh) 一种日志分发方法及装置
CN110968474A (zh) 用于使用人工智能来管理可变状态计算设备的操作的方法和装置
WO2019085754A1 (zh) 应用清理方法、装置、存储介质及电子设备
CN114024831B (zh) 一种异常事件预警方法、装置和系统
CN115963914A (zh) 基于混合策略的智能移动设备电源管理方法及装置
CN110737322A (zh) 信息处理方法及电子设备
US20160055037A1 (en) Analysis controller, analysis control method and computer-readable medium
KR20130076332A (ko) 소모 전력을 최소화하기 위한 이동 단말기 및 방법

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: 18911491

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: 18911491

Country of ref document: EP

Kind code of ref document: A1