Data interaction method and system of electronic equipment

Abstract

The invention relates to the technical field of communication networks, in particular to a data interaction method and system for electronic equipment, and the method comprises the steps: recording interaction data of each dimension at each moment in a data transmission process; according to the interaction data of the same dimension at the local moment of each moment, the mutation degree of the interaction data of each dimension at each moment is obtained, and suspected abnormal data is screened out; the method comprises the following steps: acquiring a reference value of each moment according to a sudden change degree of interaction data of different dimensions at each moment, acquiring a single-dimensional abnormal degree of suspected abnormal data in combination with interaction data of a dimension corresponding to the suspected abnormal data at a local moment of a corresponding moment, and acquiring an abnormal degree of each moment in combination with a time sequence distance between moments; and the real abnormity in the data interaction process is identified. According to the invention, by analyzing the interaction data of each dimension at each moment in the interaction process, the real abnormity in the data interaction process is accurately monitored, so that the security of data interaction is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication network, and in particular to a data interaction method and system of electronic equipment. BACKGROUND

[0002] 5G network data interaction has the advantages of high bandwidth and low latency, and using 5G network can make the system faster and more stable for data interaction. However, in the process of data interaction of 5G tablet computer, abnormality may occur due to actual failure or performance degradation in the device, network or data transmission link. The abnormality in the process of data interaction will directly affect the stability of the system and the user experience. In order to protect the user experience and service quality of 5G tablet computer in the process of data interaction, and improve the security and reliability of data interaction, it is necessary to monitor whether abnormality occurs in the process of data interaction of 5G tablet computer.

[0003] The traditional method of monitoring whether abnormality occurs in the process of data interaction of 5G tablet computer mainly analyzes whether the interaction data of each dimension (such as packet loss rate, bandwidth utilization rate, etc.) changes suddenly in the process of data interaction. However, in the process of data interaction of 5G tablet computer, the 5G tablet computer may switch between different base stations due to the movement of the device, and the 5G tablet computer may switch from one normal working mode to another normal working mode due to the user's operation. Base station switching or working mode switching will cause the interaction data of each dimension to change suddenly, so it is not possible to accurately monitor whether abnormality occurs in the process of data interaction of 5G tablet computer. SUMMARY

[0004] The present application provides a data interaction method and system of electronic equipment to solve the existing problem that the traditional method of monitoring whether abnormality occurs in the process of data interaction cannot accurately monitor whether abnormality occurs in the process of data interaction of electronic equipment.

[0005] The data interaction method and system of electronic equipment of the present application adopt the following technical solutions: One embodiment of the present application provides a data interaction method of electronic equipment, which comprises the following steps: record the interaction data of each dimension at each time in the data transmission process; According to the interaction data of the same dimension at each local time, the degree of change of the interaction data of each dimension at each time is obtained, and then the suspected abnormal data is selected from the interaction data of all dimensions at all times. According to the mutation degree of the interaction data of different dimensions at each moment, a reference value of each moment is obtained, a trend abnormality degree of the suspected abnormal data is obtained in combination with the trend of the corresponding dimension of the suspected abnormal data at the corresponding moment, an amplitude abnormality degree of the suspected abnormal data is obtained according to the reference value of each moment and the amplitude of the corresponding dimension of the suspected abnormal data at the corresponding moment, and a single-dimension abnormality degree of the suspected abnormal data is obtained in combination with the trend abnormality degree of the suspected abnormal data. According to the single-dimension abnormality degree of all suspected abnormal data at each moment, an abnormality degree of each moment is obtained in combination with the time sequence distance between the single moment and each local moment, and a real abnormality in the data interaction process is further identified.

[0006] Preferably, the method for obtaining the mutation degree of the interaction data of each dimension at each moment according to the interaction data of the same dimension at each local moment of each moment comprises the following specific method: A local time range is preset For any moment, the moments before the moment are taken as the moments before the moment, and the moments after the moment are taken as the moments after the moment, and the moments before the moment and the moments after the moment are taken as the local moments of the moment. For the interaction data of any dimension at the first local moment of the moment, the absolute value of the difference between the interaction data of the dimension at the first local moment of the moment and the interaction data of the dimension at the moment is divided by the time sequence distance between the moment and the first local moment of the moment, and the obtained ratio is taken as the difference factor of the interaction data of the dimension at the first local moment of the moment. The mean value of the mutation factors of the interaction data of the dimension at all local moments of the moment is divided by the standard deviation of the interaction data of the dimension at all local moments of the moment, and the obtained ratio is taken as the difference degree of the interaction data of the dimension at the moment. The difference degree of the interaction data of the dimension at the moment is normalized, and the normalized result is taken as the mutation degree of the interaction data of the dimension at the moment. Preferably, the method for obtaining the reference value of each moment according to the mutation degree of the interaction data of different dimensions at each moment comprises the following specific method: For the interaction data of the first dimension and the second dimension at any moment, the difference between the interaction data of the first dimension at the moment and the interaction data of the first dimension at the moment is divided by the time sequence distance between the moment and the moment, and the obtained ratio is taken as the reference value of the moment.

[0007] Preferably, the method for obtaining the reference value of each moment according to the mutation degree of the interaction data of different dimensions at each moment comprises the following specific method: For the interaction data of the first dimension and the second dimension at any moment, the difference between the interaction data of the first dimension at the moment and the interaction data of the first dimension at the moment is divided by the time sequence distance between the moment and the moment, and the obtained ratio is taken as the reference value of the moment. ​​​​​​The degree of mutation in the interaction data of the first dimension and the second dimension The product of the mutation rates of the interaction data in each dimension is linearly normalized, and the result of the linear normalization is used as the result of the first step at the given time. The dimension and the first The degree of co-mutation among dimensions; Based on the degree of co-mutation among all different dimensions at the given time, a symmetric matrix of the interaction data of all dimensions at the given time is constructed as the co-mutation matrix of all dimensions at the given time. Eigenvalue decomposition is performed on the co-mutation matrix of all dimensions at the stated time to obtain several eigenvalues ​​of the co-mutation matrix of all dimensions at the stated time. The ratio of the largest eigenvalue to the smallest eigenvalue of the co-mutation matrix of all dimensions at the stated time is used as the reference value for the stated time.

[0008] Preferably, the specific method for obtaining the trend anomaly degree of the suspected abnormal data includes: In the formula, Indicates the degree of trend abnormality in any suspected outlier data; The standard deviation of the amplitude of the interactive data of the target dimension at all previous local time points corresponding to the suspected abnormal data is represented. The standard deviation of the amplitude of the interactive data of the target dimension at all subsequent local time points corresponding to the suspected abnormal data is represented. This indicates the number of preceding local moments and the number of following local moments corresponding to the suspected abnormal data. Indicates the time corresponding to the suspected abnormal data. The changing trend of the target dimension at a given local moment; Indicates the time corresponding to the suspected abnormal data. The changing trend of the target dimension at a local time after the event; Indicates the time corresponding to the suspected abnormal data. Reference weights for each previous local time step; Indicates the time corresponding to the suspected abnormal data. Reference weights at each subsequent local time step; Indicates the time corresponding to the suspected abnormal data. The magnitude of the target dimension at each local time point; This represents the mean of the magnitude of the target dimension at all previous local time points corresponding to the suspected abnormal data. Indicates the time corresponding to the suspected abnormal data. The magnitude of the target dimension at a local time after the event; a mean value of amplitudes of the target dimension at all post-local time points corresponding to the time point of the suspected abnormal data; represents an absolute value function.

[0009] Preferably, the method for obtaining the amplitude abnormality degree of the suspected abnormal data according to the reference value of each time point and the amplitude of the dimension corresponding to the suspected abnormal data at the local time point of the corresponding time point comprises the following specific method: For any suspected abnormal data, the dimension corresponding to the abnormal data is denoted as the target dimension, the interaction data of the target dimension at all pre-local time points and all post-local time points corresponding to the time point of the suspected abnormal data are respectively subjected to STL decomposition, and the trend item of the target dimension at all pre-local time points and the trend item of the target dimension at all post-local time points corresponding to the time point of the suspected abnormal data are respectively obtained. The trend items of the target dimension at all pre-local time points and the trend items of the target dimension at all post-local time points are subjected to DTW matching, and a plurality of matching pairs of the suspected abnormal data are obtained. According to each matching pair of the suspected abnormal data and the reference value of each time point, the amplitude abnormality degree of the suspected abnormal data is obtained.

[0010] Preferably, the method for obtaining the amplitude abnormality degree of the suspected abnormal data according to the reference value of each time point and the reference value of each time point of the suspected abnormal data comprises the following specific method: For any matching pair of the suspected abnormal data, the sum of the reference values of the two elements in the matching pair of the suspected abnormal data is subjected to weight normalization, and the obtained weight normalization result is taken as the reference weight of the matching pair of the suspected abnormal data. The absolute value of the difference in amplitude between the two elements in the matching pair of the suspected abnormal data is taken as the amplitude distance of the two elements in the matching pair of the suspected abnormal data, and the square root of the sum of the square of the amplitude distance and the time sequence distance of the two elements in the matching pair of the suspected abnormal data is taken as the matching distance of the two elements in the matching pair of the suspected abnormal data. The product of the matching distance of the two elements in the matching pair of the suspected abnormal data and the reference weight of the matching pair of the suspected abnormal data is taken as the amplitude abnormality factor of the matching pair of the suspected abnormal data. The cumulative sum of the amplitude abnormality factors of all matching pairs of the suspected abnormal data is subjected to negative correlation normalization, and the obtained negative correlation normalization result is taken as the amplitude abnormality degree of the suspected abnormal data.

[0011] Preferably, the method for obtaining the single-dimensional abnormality degree of the suspected abnormal data comprises the following specific method: For any suspected abnormal data, the product of the amplitude abnormality degree and the trend abnormality degree of the suspected abnormal data is taken as the single-dimensional abnormality degree of the suspected abnormal data.

[0012] Preferably, the abnormal degree of each time point is obtained according to the single time point and the single dimension abnormal degree of all suspected abnormal data at each local time point of the single time point, and the time sequence distance between the single time point and each local time point, and the specific method comprises the following steps: In the formula, represents the abnormal degree of any time point; represents the number of local time points of the time point; represents the primary abnormal degree of the time point; represents the primary abnormal degree of the first local time point of the time point; represents the time sequence distance between the time point and the first local time point thereof; represents the absolute value function; represents the sigmoid function. Preferably, the specific method for identifying the real abnormality in the data interaction process comprises the following steps:

[0013] Preferably, the specific method for identifying the real abnormality in the data interaction process comprises the following steps:

[0013] If the abnormal degree of the time point is greater than or equal to the abnormal degree threshold value, a real abnormality occurs in the data interaction process at the time point. Another embodiment of the present application provides a data interaction system of an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of any one of the data interaction methods of the electronic device.

[0014] The technical scheme of the present application has the following beneficial effects: the present application collects the interaction data of each dimension in the data interaction process of the 5G tablet computer, further obtains the mutation degree of the interaction data of each dimension at each time point according to the interaction data of the same dimension at each local time point of each time point, and then screens out suspected abnormal data from all the interaction data of all the time points and all the dimensions; the suspected abnormal data includes real abnormality and pseudo abnormality caused by base station switching or working mode switching, and the suspected abnormal data is analyzed by focusing to quickly monitor the abnormality in the data interaction process of the 5G tablet computer.

[0015] The technical scheme of the present application has the following beneficial effects: the present application collects the interaction data of each dimension in the data interaction process of the 5G tablet computer, further obtains the mutation degree of the interaction data of each dimension at each time point according to the interaction data of the same dimension at each local time point of each time point, and then screens out suspected abnormal data from all the interaction data of all the time points and all the dimensions; the suspected abnormal data includes real abnormality and pseudo abnormality caused by base station switching or working mode switching, and the suspected abnormal data is analyzed by focusing to quickly monitor the abnormality in the data interaction process of the 5G tablet computer.

[0016] ​It needs to be further explained that when an abnormality occurs in the data interaction process of the 5G tablet computer, the interaction data of each dimension will mutate, and the external conditions before and after the abnormality in the data interaction process do not change, and the interaction data of each dimension before mutation is similar to the interaction data of each dimension after mutation; when the working mode of the 5G tablet computer is switched, the scheduling strategy and bandwidth utilization mode of the 5G tablet computer will be adjusted, and the trend of the interaction data of each dimension before and after mutation is not similar; when the 5G tablet computer switches between different base stations, the physical link is substantially changed, and the interaction data of each dimension will have a step change in amplitude at the switching moment, and the system will also be re-stabilized after switching, so the single-dimensional abnormality degree of the suspected abnormal data is obtained, and finally the abnormality degree of the moment is obtained by combining the single-dimensional abnormality degree of each dimension at the same moment, so as to accurately monitor whether an abnormality occurs in the data interaction process of the 5G tablet computer, and to ensure the user experience and service quality in the data interaction process of the 5G tablet computer, and to improve the security and reliability of data interaction. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, below the drawings needed to be used in the embodiment or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0018] Figure 1 The step flow chart of the data interaction method of the electronic device. DETAILED DESCRIPTION

[0019] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined object, the specific implementation, structure, features and effects of the data interaction method and system of an electronic device according to the present application are described in detail as follows. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0021] The specific scheme of the data interaction method and system of an electronic device provided by the present application is described in detail below with reference to the drawings.

[0022] Please refer toFigure 1 which shows a step flow chart of a data interaction method of an electronic device provided by an embodiment of the present application, and the method comprises the following steps: Step S001: record the interaction data of each dimension at each moment in the data transmission process.

[0023] It should be noted that in the data interaction process of the 5G tablet computer, anomalies may occur due to actual faults or performance degradation in the device, network or data transmission link. The anomalies in the data interaction process will directly affect the stability of the system and the user experience. Therefore, the present embodiment proposes a data interaction method for a 5G tablet computer, which specifically analyzes the interaction data of each dimension at each moment to monitor whether an anomaly occurs in the data interaction process of the 5G tablet computer, thereby ensuring user experience and service quality in the data interaction process of the 5G tablet computer and improving the security and reliability of data interaction.

[0024] Specifically, the interaction data of each dimension at each moment is recorded in the data interaction process of the 5G tablet computer. The dimensions of the interaction data include but are not limited to packet loss rate, bandwidth utilization, etc. In the present embodiment, 0.1 seconds are taken as a moment for description.

[0025] Step S002: according to the interaction data of the same dimension at each local moment of each moment, obtain the mutation degree of the interaction data of each dimension at each moment, and then screen out suspected abnormal data from all the interaction data of all dimensions at all moments.

[0026] It should be noted that when an anomaly occurs in the data interaction process of the 5G tablet computer, the interaction data of each dimension will be affected by the anomaly and will mutate. However, in the data interaction process of the 5G tablet computer, the 5G tablet computer may switch between different base stations due to the movement of the device, and the 5G tablet computer may switch from one normal working mode to another normal working mode (for example, from the working mode of "browsing webpages" to the working mode of "downloading files") due to the user's operation. Whether the 5G tablet computer switches between different base stations or the 5G tablet computer switches from one normal working mode to another normal working mode, the interaction data of each dimension will mutate. Therefore, the present embodiment obtains the mutation degree of the interaction data of each dimension at each moment by the interaction data of the same dimension at each local moment of each moment, and then screens out suspected abnormal data from all the interaction data of all dimensions at all moments; the suspected abnormal data includes real anomalies and pseudo anomalies caused by base station switching or working mode switching. Subsequently, the suspected abnormal data is analyzed by focusing to quickly monitor the anomalies occurring in the data interaction process of the 5G tablet computer.

[0027] Preferably, in a specific embodiment of the present invention, a local time range is preset. The The specific scope can be set according to the actual situation. This embodiment does not make a rigid requirement. In this embodiment, it is used as... Taking this as an example, for any given time, the time before... Each moment is taken as a previous local moment, and the next moment is taken as a subsequent local moment. Each moment is designated as a subsequent local moment of the stated moment, and the preceding and subsequent local moments of the stated moment are denoted as local moments of the stated moment. For the time mentioned above, the first Interaction data of any dimension at a local time; the first time of the aforementioned time. The absolute value of the difference between the interaction data of the dimension at a local time point and the interaction data of the dimension at the specified time point, compared with the value of the interaction data of the dimension at the specified time point and the specified time point. The temporal distance between local moments is used as the ratio of the obtained values ​​for the first time point. The difference factor of the interaction data of the aforementioned dimension at each local time point; Furthermore, the mean of the mutation factors of the interaction data of the dimension at all local time points at the stated time is compared with the standard deviation of the interaction data of the dimension at all local time points at the stated time. The ratio obtained is used as the degree of difference of the interaction data of the dimension at the stated time. The degree of difference of the interaction data of the dimension at the stated time is normalized (normalization can be performed using the sigmoid function). The normalization result is used as the degree of mutation of the interaction data of the dimension at the stated time.

[0028] As an example, the specific formula for calculating the degree of mutation of the interaction data of the specified dimension at the specified time is as follows: In the formula, This indicates the degree of mutation in the interactive data of the specified dimension at the specified time. This indicates the number of local moments at the given moment; This represents the magnitude of the interactive data in the dimension at the given time. Represents the first time. The magnitude of the interactive data of the aforementioned dimension at a local time point; Indicates the time and its first The temporal distance between local moments; The standard deviation of the interaction data in the dimension at all local time points at the given time is represented. This represents the function that takes the absolute value. This represents the sigmoid function, which is used for normalization in this embodiment.

[0029] It should be noted that the greater the difference between the time and its local time in the interaction data of a certain dimension, the greater the change in the interaction data of the time in the dimension, and the shorter the time taken for the interaction data of the dimension to change at the time, the more severe the change in the interaction data of the dimension at the time. Therefore, a negative correlation calculation weight is given to the time sequence distance between the time and its local time. At the same time, in order to avoid the influence of the strong volatility of the dimension itself under normal circumstances, the standard deviation of the amplitude of the interaction data of the dimension at all local times of the time is given a negative correlation calculation weight, so as to accurately evaluate the mutation degree of the interaction data of the dimension at the time.

[0030] It should be further noted that since the interaction data of each dimension will mutate whether it is a real anomaly in the process of data interaction or the 5G tablet computer switches between different base stations or the working mode switches, the mutation degree of the interaction data of each dimension at each time can be used to screen out suspected abnormal data from the interaction data of all dimensions at all times. Subsequently, the suspected abnormal data is analyzed by focusing to quickly monitor the anomalies in the process of data interaction of the 5G tablet computer.

[0031] Specifically, a mutation degree threshold is preset The specific value of the may be set by itself according to actual conditions, and the present embodiment does not make a hard requirement. In the present embodiment, the is taken as an example for description; for the interaction data of any dimension at any time, if the mutation degree of the interaction data of the dimension at the time is greater than or equal to , the interaction data of the dimension at the time is recorded as suspected abnormal data.

[0032] At this point, all suspected abnormal data at all times is obtained.

[0033] Step S003: According to the mutation degree of the interaction data of different dimensions at each time, the reference value of each time is obtained, the trend abnormal degree of the suspected abnormal data is obtained by combining the trend of the corresponding dimension of the suspected abnormal data at the corresponding local time of the corresponding time, the amplitude abnormal degree of the suspected abnormal data is obtained according to the reference value of each time and the amplitude of the corresponding dimension of the suspected abnormal data at the corresponding local time of the corresponding time, and the single-dimensional abnormal degree of the suspected abnormal data is obtained by combining the trend abnormal degree of the suspected abnormal data.

[0034] It should be noted that when an abnormality occurs in the data interaction process of the 5G tablet computer, the interaction data of each dimension will mutate, and the external conditions before and after the abnormality in the data interaction process do not change (for example, base station load, network environment), the interaction data of each dimension before mutation is similar to the interaction data of each dimension after mutation, and when the working mode of the 5G tablet computer is switched, the scheduling strategy and bandwidth utilization mode inside the 5G tablet computer will be adjusted, thereby gradually changing the evolution direction of the interaction data in the subsequent period of time, for example, the throughput presents a continuous upward trend, the delay presents a continuous downward trend, and the bandwidth utilization rate gradually increases, and the trends of the interaction data of each dimension before and after mutation are not similar. When the 5G tablet computer switches between different base stations, the physical link is substantially changed, and due to the differences in coverage environment, signal strength and scheduling resources of the new and old base stations, the interaction data of each dimension will have a step change in amplitude at the moment of switching, and the system will also stabilize after switching, that is, the amplitudes of the interaction data of each dimension before and after mutation are different. Therefore, when the amplitudes and trends of the interaction data of each dimension before and after mutation are similar, the mutation is more likely to be a real abnormality.

[0035] It should be further noted that when multiple dimensions of interaction data mutate at the same time in the data interaction process of the 5G tablet computer, it indicates that a physical factor (such as a strong interference electromagnetic field or a momentary overload of the core processor) simultaneously affects multiple subsystems of the communication system at the same time, thereby causing multiple dimensions of interaction data to mutate at the same time. The mutation of multiple dimensions of interaction data at the same time indicates that the overall working state of the system has changed, thereby making the interaction data of each dimension at the same time have higher information value. In order to accurately distinguish between real abnormalities and pseudo abnormalities caused by base station switching or working mode switching, the reference value of each time is obtained by the mutation degree of multiple dimensions of interaction data at the same time, and real abnormalities and pseudo abnormalities in the data interaction process of the 5G tablet computer are accurately distinguished.

[0036] Preferably, in a specific embodiment of the present application, for the interaction data of the first dimension and the second dimension at any time, the product of the mutation degree of the interaction data of the first dimension and the mutation degree of the interaction data of the second dimension at the time is linearly normalized (the normalization range of the linear normalization is the product of the mutation degrees of the interaction data of all different dimensions at the time), and the result of the linear normalization is taken as the cooperative mutation degree between the first dimension and the second dimension at the time.

[0037] ​​​​​​According to the degree of the coordinated mutation between all the different dimensions at the moment, a symmetric matrix of the interaction data of all the dimensions at the moment is constructed. Since the construction of the symmetric matrix is a known prior art, it will not be described here. The symmetric matrix is the coordinated mutation matrix of all the dimensions at the moment, as shown below: wherein, represents the degree of the coordinated mutation between the first dimension and the first dimension at the moment; represents the degree of the coordinated mutation between the first dimension and the second dimension at the moment; represents the degree of the coordinated mutation between the second dimension and the first dimension at the moment; represents the degree of the coordinated mutation between the second dimension and the second dimension at the moment; represents the degree of the coordinated mutation between the first dimension and the dimension at the moment; represents the degree of the coordinated mutation between the dimension and the first dimension at the moment; represents the degree of the coordinated mutation between the dimension and the dimension at the moment; represents the number of dimensions.

[0038] Further, the coordinated mutation matrix of all the dimensions at the moment is subjected to eigenvalue decomposition to obtain several eigenvalues of the coordinated mutation matrix of all the dimensions at the moment. The ratio of the maximum eigenvalue to the minimum eigenvalue of the coordinated mutation matrix of all the dimensions at the moment is taken as the reference value of the moment.

[0039] It should be noted that the greater the degree of the coordinated mutation between different dimensions at the moment, the stronger the mutation of the two dimensions at the moment. The maximum eigenvalue and the minimum eigenvalue of the coordinated mutation matrix represent the strongest coordination mode and the weakest coordination mode of the coordinated mutation matrix, respectively. The greater the ratio of the maximum eigenvalue to the minimum eigenvalue of the coordinated mutation matrix of all the dimensions at the moment, the more information of the coordinated matrix is concentrated on a dominant mode, which corresponds to the real global abnormal scenario of the synchronous and coordinated strong mutation of the interaction data of all the dimensions, indicating that the overall working state of the system has undergone a key change at the moment, so that the interaction data of each dimension at the moment has higher information value.

[0040] It should be further explained that when an anomaly occurs during the data interaction process of a 5G tablet, the interaction data in each dimension before the mutation is similar to the interaction data in each dimension after the mutation; however, when the working mode of the 5G tablet switches, the trends of the interaction data in each dimension before and after the mutation are not similar. Therefore, this can be used as a basis, combined with the reference value at each moment, to obtain the degree of trend anomaly of the suspected abnormal data, in order to distinguish whether the suspected abnormal data is a real anomaly or a pseudo-anomaly caused by the switching of working modes.

[0041] Preferably, in a specific embodiment of the present invention, for any local moment corresponding to any suspected abnormal data, the reference value of the local moment corresponding to the suspected abnormal data is weighted and normalized (the range of the weight normalization is the reference value of all local moments corresponding to the suspected abnormal data), and the result of the weight normalization is used as the reference weight of the local moment corresponding to the suspected abnormal data. Furthermore, the dimension corresponding to the abnormal data is denoted as the target dimension. Based on the amplitude of the interaction data of the target dimension at each local time point corresponding to the suspected abnormal data, and combined with the reference weights at each local time point corresponding to the suspected abnormal data, the trend abnormality degree of the suspected abnormal data is obtained. The specific calculation formula is as follows: In the formula, This indicates the degree of abnormality in the trend of the suspected abnormal data; The standard deviation of the amplitude of the interactive data of the target dimension at all previous local time points corresponding to the suspected abnormal data is represented. The standard deviation of the amplitude of the interactive data of the target dimension at all subsequent local time points corresponding to the suspected abnormal data is represented. This indicates the number of preceding local moments and the number of following local moments corresponding to the suspected abnormal data. The time corresponding to the suspected abnormal data is the first... The changing trend of the target dimension at a given local moment; The time corresponding to the suspected abnormal data is the first... The changing trend of the target dimension at a local time after the event; The time corresponding to the suspected abnormal data is the first... Reference weights for each previous local time step; The time corresponding to the suspected abnormal data is the first... Reference weights at each subsequent local time step; The time corresponding to the suspected abnormal data is the first... The magnitude of the target dimension at each local time point; This represents the mean of the magnitude of the target dimension at all previous local time points corresponding to the suspected abnormal data. The time corresponding to the suspected abnormal data is the first... The magnitude of the target dimension at a local time after the event; This represents the mean of the magnitude of the target dimension at all subsequent local time points corresponding to the suspected abnormal data. This represents the function that takes the absolute value.

[0042] It should be noted that, This indicates the time corresponding to the suspected abnormal data. The degree to which the interaction data of the target dimension deviates from its overall trend at a given local time point. This indicates the time corresponding to the suspected abnormal data. This embodiment analyzes the deviation of the target dimension's interaction data from its overall trend at each subsequent local time point, and also analyzes the degree of deviation of the target dimension's interaction data from its overall trend at each preceding and subsequent local time points, quantifying the trend similarity of the target dimension's interaction data between preceding and subsequent local time points. However, since the reference value of different time points varies, direct comparison may lead to distorted results. Therefore, this embodiment further introduces the reference value of local time points as a weight to weight the trend differences of each local time point. This can highlight the contribution of time points with higher reference value to the overall trend, thereby more accurately quantifying the trend similarity characteristics of the target dimension's interaction data before and after the mutation under real anomalies.

[0043] It should be further explained that when an anomaly occurs during the data interaction process of a 5G tablet, the interaction data in each dimension before the mutation is similar to the interaction data in each dimension after the mutation; however, when the 5G tablet switches between different base stations, the amplitude of the interaction data in each dimension before and after the mutation is different. Therefore, this can be used as a basis, combined with the reference value at each moment, to obtain the degree of amplitude anomaly of the suspected abnormal data, in order to distinguish whether the suspected abnormal data is a real anomaly or a pseudo-anomaly caused by base station switching.

[0044] Preferably, in a specific embodiment of the present application, for any suspected abnormal data, the dimension corresponding to the abnormal data is recorded as the target dimension, the interaction data of the target dimension at all previous local time points and all subsequent local time points of the time point corresponding to the suspected abnormal data are respectively subjected to STL (Seasonal and Trend decomposition using Loess) decomposition, and the trend items of the target dimension at all previous local time points and all subsequent local time points of the time point corresponding to the suspected abnormal data are respectively obtained. The trend items of the target dimension at all previous local time points and all subsequent local time points of the time point corresponding to the suspected abnormal data are subjected to DTW (Dynamic Time Warping) matching, and a plurality of matching pairs of the suspected abnormal data are obtained. Since the STL decomposition method and the DTW matching method are both prior art, they will not be described in detail in this embodiment. Further, for any matching pair of the suspected abnormal data, the sum of the reference values at the time points corresponding to the two elements of the matching pair of the suspected abnormal data is subjected to weight normalization (the range of the weight normalization is the sum of the reference values at the time points corresponding to the two elements of all matching pairs of suspected abnormal data), and the obtained weight normalization result is taken as the reference weight of the matching pair of the suspected abnormal data. The absolute value of the difference in amplitude between the two elements in the matching pair of the suspected abnormal data is taken as the amplitude distance of the two elements in the matching pair of the suspected abnormal data, and the square root of the sum of the squares of the amplitude distance and the time sequence distance of the two elements in the matching pair of the suspected abnormal data is taken as the matching distance of the two elements in the matching pair of the suspected abnormal data. The product of the matching distance of the two elements in the matching pair of the suspected abnormal data and the reference weight of the matching pair of the suspected abnormal data is taken as the amplitude anomaly factor of the matching pair of the suspected abnormal data. The cumulative sum of the amplitude anomaly factors of all matching pairs of the suspected abnormal data is subjected to negative correlation normalization (the negative correlation normalization can be performed by using the function, where a is a natural constant, and b is a constant greater than 0), and the obtained negative correlation normalization result is taken as the amplitude abnormality degree of the suspected abnormal data. represents an exponential function with a natural constant as the base number, is the input of the model, and the implementer can set the inverse proportional function and the normalization function according to the actual situation), and the obtained negative correlation normalization result is taken as the amplitude abnormality degree of the suspected abnormal data.

[0045] As an example, the specific calculation formula of the amplitude abnormality degree of the suspected abnormal data is as follows: In the formula, a is a natural constant, b is a constant greater than 0, and c is a constant greater than 0. represents the amplitude abnormality degree of the suspected abnormal data. ​a number of matching pairs representing the suspected abnormal data; a reference weight of the i-th matching pair representing the suspected abnormal data; a reference weight of the i-th matching pair representing the suspected abnormal data; a reference weight of the i-th matching pair representing the suspected abnormal data; a time distance between two elements in the i-th matching pair representing the suspected abnormal data; a time distance between two elements in the i-th matching pair representing the suspected abnormal data; a time distance between two elements in the i-th matching pair representing the suspected abnormal data; an exponential function with a natural constant as a base.

[0046] It should be noted that the interaction data corresponding to the dimensions before and after the mutation caused by the base station switching is similar in trend, but the overall amplitude is not similar, while the overall amplitude before and after the mutation of the real abnormality is also similar, so the embodiment decomposes the STL, only matches the trend item by DTW, thereby avoiding the influence of the trend, and then accurately quantifies the similarity of the overall amplitude before and after the mutation of the suspected abnormal data, so as to accurately distinguish the real abnormality from the pseudo abnormality caused by the base station switching. After obtaining the amplitude abnormality factor and the trend abnormality factor of the suspected abnormal data, the single-dimensional abnormality factor of the suspected abnormal data can be obtained according to the amplitude abnormality factor and the trend abnormality factor of the suspected abnormal data.

[0047] Specifically, for any suspected abnormal data, the product of the amplitude abnormality degree and the trend abnormality degree of the suspected abnormal data is taken as the single-dimensional abnormality degree of the suspected abnormal data.

[0048] It should be noted that the mutation caused by the real abnormality in the data interaction process of the 5G tablet computer is similar in amplitude and trend before and after the mutation, so the greater the amplitude abnormality degree and the trend abnormality degree of the suspected abnormal data, the more likely the suspected abnormal data is a real abnormality.

[0049] At this point, the single-dimensional abnormality degree of the suspected abnormal data is obtained.

[0050] Step S004: According to the single time and the single-dimensional abnormality degree of all suspected abnormal data at each local time, and combining the time distance between the single time and each local time, the abnormality degree of each time is obtained, and then the real abnormality in the data interaction process is identified.

[0051] It should be noted that when a real abnormal situation occurs in the data interaction process of the 5G tablet, the real abnormality will cause the interaction data of each dimension to be abnormal in a short time. After obtaining the single-dimensional abnormality degree of the suspected abnormal data in step S003, the abnormality degree at the moment can be obtained according to the single-dimensional abnormality degree of each dimension at the moment, so as to accurately monitor whether an abnormality occurs in the data interaction process of the 5G tablet and to ensure the user experience and service quality in the data interaction process of the 5G tablet and improve the security and reliability of data interaction.

[0052] Preferably, in a specific embodiment of the present application, for any moment, the sum of the single-dimensional abnormality degrees of all suspected abnormal data at the moment is taken as the primary abnormality degree at the moment (if there is no suspected abnormal data at the moment, the primary abnormality degree at the moment is 0); Further, the primary abnormality degree of each local moment at the moment is obtained, and the abnormality degree at the moment is obtained according to the moment and the primary abnormality degree of each local moment thereof, in combination with the time sequence distance between the moment and each local moment thereof. The specific calculation formula is as follows: In the formula, represents the abnormality degree at the moment; represents the number of local moments at the moment; represents the primary abnormality degree at the moment; represents the primary abnormality degree of the first local moment at the moment; represents the time sequence distance between the moment and the first local moment thereof; represents the absolute value function; represents the sigmoid function, which is used for normalization in the present embodiment.

[0053] It should be noted that the abnormality degree at the moment represents the possibility of real abnormality at the moment. The primary abnormality degree at the moment is the sum of the single-dimensional abnormality factors of all suspected abnormal data at the moment. Since the interaction data abnormality of each dimension caused by real abnormality bursts in a short time, the greater the single-dimensional abnormality factors of multiple dimensions at the same moment, the more likely it is a real fault. Meanwhile, the greater the difference between the primary abnormality degree at the moment and that of the surrounding moment, the more concentrated the interaction data abnormality of each dimension in a short time, and the more likely it is that a real abnormal situation occurs at the moment. After obtaining the abnormality degree at the moment, the real abnormality in the data interaction process of the 5G tablet can be accurately monitored according to the abnormality degree at the moment, so as to ensure the user experience and service quality in the data interaction process of the 5G tablet and improve the security and reliability of data interaction.

[0054] Specifically, an abnormality degree threshold value at a time point is preset The specific value of the abnormality degree threshold value can be set by combining with the actual situation, and the embodiment does not make a hard requirement, and in the embodiment, the abnormality degree threshold value is taken as an example for description, for any time point, if the abnormality degree of the time point is greater than or equal to the abnormality degree threshold value, a real abnormality occurs in the process of data interaction at the time point.

[0055] Another embodiment of the present application provides a data interaction system of an electronic device, comprising a memory, a processor and a computer program stored in the memory and capable of running on the processor, and the processor implements the data interaction method of the electronic device in steps S001 to S004 when the computer program is executed.

[0056] The above only describes the preferred embodiments of the present application and is not intended to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.​​​

Claims

1. A data interaction method for an electronic device, characterized in that, The method includes the following steps: Record the interactive data of each dimension at each moment during the data transmission process; Based on the interaction data of the same dimension at each local time, the degree of mutation of the interaction data of each dimension at each time is obtained, and then suspected abnormal data is screened out from the interaction data of all dimensions at all times. Based on the degree of mutation of the interactive data in different dimensions at each time point, the reference value of each time point is obtained. Combined with the trend of the corresponding dimension of the suspected abnormal data at the corresponding time point, the trend abnormality of the suspected abnormal data is obtained. Based on the reference value of each time point and the amplitude of the corresponding dimension of the suspected abnormal data at the corresponding time point, the amplitude abnormality of the suspected abnormal data is obtained. Combined with the trend abnormality of the suspected abnormal data, the single-dimensional abnormality of the suspected abnormal data is obtained. Based on the unidimensional anomaly degree of all suspected abnormal data at a single time point and at each of its local time points, and combined with the temporal distance between the single time point and each local time point, the anomaly degree of each time point is obtained, thereby identifying the real anomalies in the data interaction process.

2. The data interaction method for an electronic device according to claim 1, characterized in that, The specific method for obtaining the degree of mutation of interaction data in each dimension at each time step based on the interaction data of the same dimension at each local time step includes: Preset a local time range For any given time, the time preceding the given time... Each moment is taken as a previous local moment, and the next moment is taken as a subsequent local moment. Each moment is designated as a subsequent local moment of the stated moment, and the preceding and subsequent local moments of the stated moment are denoted as local moments of the stated moment. For the time mentioned above, the first Interaction data of any dimension at a local time; the first time of the aforementioned time. The absolute value of the difference between the interaction data of the dimension at a local time point and the interaction data of the dimension at the specified time point, compared with the value of the interaction data of the dimension at the specified time point and the specified time point. The temporal distance between local moments is used as the ratio of the obtained values ​​for the first time point. The difference factor of the interaction data of the aforementioned dimension at each local time point; The mean of the mutation factors of the interaction data of the dimension at all local time points at the given time point is divided by the standard deviation of the interaction data of the dimension at all local time points at the given time point. The ratio obtained is taken as the degree of difference of the interaction data of the dimension at the given time point. The degree of difference of the interaction data of the dimension at the given time point is normalized, and the normalization result is taken as the degree of mutation of the interaction data of the dimension at the given time point.

3. The data interaction method for an electronic device according to claim 1, characterized in that, The method for obtaining the reference value at each moment based on the degree of mutation of interaction data in different dimensions at each moment includes: For any time, the... The dimension and the first Interaction data in one dimension, for the time specified. The degree of mutation in the interaction data of the first dimension and the second dimension The product of the mutation rates of the interaction data in each dimension is linearly normalized, and the result of the linear normalization is used as the result of the first step at the given time. The dimension and the first The degree of co-mutation among dimensions; Based on the degree of co-mutation among all different dimensions at the given time, a symmetric matrix of the interaction data of all dimensions at the given time is constructed as the co-mutation matrix of all dimensions at the given time. Eigenvalue decomposition is performed on the co-mutation matrix of all dimensions at the stated time to obtain several eigenvalues ​​of the co-mutation matrix of all dimensions at the stated time. The ratio of the largest eigenvalue to the smallest eigenvalue of the co-mutation matrix of all dimensions at the stated time is used as the reference value for the stated time.

4. The data interaction method for an electronic device according to claim 2, characterized in that, The specific methods for obtaining the trend anomaly degree of suspected abnormal data include: In the formula, Indicates the degree of trend abnormality in any suspected outlier data; The standard deviation of the amplitude of the interactive data of the target dimension at all previous local time points corresponding to the suspected abnormal data is represented. The standard deviation of the amplitude of the interactive data of the target dimension at all subsequent local time points corresponding to the suspected abnormal data is represented. This indicates the number of preceding local moments and the number of following local moments corresponding to the suspected abnormal data. Indicates the time corresponding to the suspected abnormal data. The changing trend of the target dimension at a given local moment; Indicates the time corresponding to the suspected abnormal data. The changing trend of the target dimension at a local time after the event; Indicates the time corresponding to the suspected abnormal data. Reference weights for each previous local time step; Indicates the time corresponding to the suspected abnormal data. Reference weights at each subsequent local time step; Indicates the time corresponding to the suspected abnormal data. The magnitude of the target dimension at each local time point; This represents the mean of the magnitude of the target dimension at all previous local time points corresponding to the suspected abnormal data. Indicates the time corresponding to the suspected abnormal data. The magnitude of the target dimension at a local time after the event; This represents the mean of the magnitude of the target dimension at all subsequent local time points corresponding to the suspected abnormal data. This represents the function that takes the absolute value.

5. The data interaction method for an electronic device according to claim 2, characterized in that, The method for obtaining the degree of amplitude abnormality of suspected abnormal data based on the reference value of each moment and the amplitude of the corresponding dimension of the suspected abnormal data at the local moment of the corresponding moment includes the following specific methods: For any suspected anomalous data, the dimension corresponding to the anomalous data is denoted as the target dimension. STL decomposition is performed on the interaction data of the target dimension at all previous local time points and all subsequent local time points corresponding to the suspected anomalous data, respectively, to obtain the trend terms of the target dimension at all previous local time points and all subsequent local time points corresponding to the suspected anomalous data. DTW matching is then performed on the trend terms of the target dimension at all previous local time points and all subsequent local time points corresponding to the suspected anomalous data, to obtain several matching pairs of the suspected anomalous data. Based on the reference value of each matching pair of the suspected anomalous data at each time point, the magnitude anomalousness of the suspected anomalous data is obtained.

6. The data interaction method for an electronic device according to claim 5, characterized in that, The specific method for obtaining the magnitude anomaly degree of the suspected anomaly data based on each matching pair of the suspected anomaly data and the reference value at each time point includes: For any matching pair of the suspected abnormal data, the sum of the reference values ​​of the two elements at the corresponding time in the matching pair of the suspected abnormal data is weighted and normalized, and the obtained weight normalization result is used as the reference weight of the matching pair of the suspected abnormal data. The absolute value of the difference in amplitude between the two elements in the matching pair of the suspected abnormal data is taken as the amplitude distance between the two elements in the matching pair of the suspected abnormal data. The square root of the sum of the squares of the amplitude distance and the temporal distance between the two elements in the matching pair of the suspected abnormal data is taken as the matching distance between the two elements in the matching pair of the suspected abnormal data. The product of the matching distance between the two elements in the matching pair of the suspected abnormal data and the reference weight of the matching pair of the suspected abnormal data is used as the amplitude anomaly factor of the matching pair of the suspected abnormal data. The sum of the amplitude anomaly factors of all matching pairs of the suspected anomalous data is negatively correlated and normalized. The negatively correlated normalization result is used as the amplitude anomaly degree of the suspected anomalous data.

7. The data interaction method for an electronic device according to claim 1, characterized in that, The specific methods for obtaining the single-dimensional anomaly degree of suspected abnormal data are as follows: For any suspected anomalous data, the product of the magnitude of the anomalousness and the trend of the suspected anomalous data is taken as the one-dimensional anomalousness of the suspected anomalous data.

8. The data interaction method for an electronic device according to claim 1, characterized in that, The method for obtaining the degree of anomaly at each time point based on the one-dimensional anomaly degree of all suspected anomaly data at a single time point and at each of its local time points, combined with the temporal distance between the single time point and each local time point, includes the following specific methods: In the formula, Indicates the degree of abnormality at any given time; This indicates the number of local time points at the given moment; This indicates the initial anomaly level at the stated time. Represents the first time. The degree of primary anomaly at a local moment; Indicates the time and its first The temporal distance between local moments; This represents the function that takes the absolute value. This represents the sigmoid function.

9. The data interaction method for an electronic device according to claim 1, characterized in that, The specific methods for identifying genuine anomalies during the data interaction process are as follows: Preset a threshold for the degree of anomaly at any given time. For any given time, if the degree of abnormality at that time is greater than or equal to... If a real anomaly occurs during the data interaction at the stated time, then...

10. A data interaction system for an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of a data interaction method for an electronic device as described in any one of claims 1-9.

Images