Communication index data prediction method and device, electronic equipment and storage medium

Abstract

The invention provides a communication index data prediction method, a communication index data prediction device, electronic equipment and a computer readable storage medium, and belongs to the technical field of wireless communication. The method comprises the following steps: acquiring first historical index data of an index to be predicted in a target cell, and second historical index data of a related index associated with the target cell; processing the first historical index data by adopting a first attention mechanism to determine first correlation data, and processing the second historical index data by adopting a second attention mechanism to determine second correlation data; determining target correlation data according to the first correlation data and the second correlation data; and processing the target correlation data through a pre-trained communication data prediction network to obtain prediction index data of the to-be-predicted index in the target cell. According to the method, the communication index data can be accurately and effectively predicted.

Description

Background Art

[0002] With the rapid development of wireless communication technology, how to accurately predict index data such as wireless performance and load is crucial for the efficient allocation of network resources, network planning, and the improvement of user experience. Existing technologies often can only predict future index data based on the historical data of the index to be predicted, with a relatively single data analysis dimension, and cannot deeply explore the complex relationships between related index data and the index data to be predicted, making it difficult to accurately predict the index data of the index to be predicted, and the prediction accuracy and prediction effect are poor.

[0003] It should be noted that the information disclosed in the above background art section is only used to enhance the understanding of the background of the present disclosure, and thus may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0004] The present disclosure provides a communication index data prediction method, a communication index data prediction device, an electronic device, and a computer-readable storage medium, thereby at least to some extent overcoming the problem of inaccurate prediction of communication index data in the prior art.

[0005] Other features and advantages of the present disclosure will become apparent through the following detailed description, or be learned in part through the practice of the present disclosure.

[0006] According to one aspect of the present disclosure, there is provided a communication index data prediction method, including: obtaining first historical index data of an index to be predicted in a target cell, and second historical index data of related indexes associated with the target cell; processing the first historical index data by using a first attention mechanism to determine first correlation data, and processing the second historical index data by using a second attention mechanism to determine second correlation data; determining target correlation data according to the first correlation data and the second correlation data; and processing the target correlation data by using a pre-trained communication data prediction network to obtain prediction index data of the index to be predicted in the target cell.

[0007] In an exemplary embodiment of the present disclosure, processing the first historical metric data using the first attention mechanism to determine first correlation data, and processing the second historical metric data using the second attention mechanism to determine second correlation data includes: performing attention calculation and feed-forward neural network processing on the first historical metric data using a first encoder to obtain first intermediate data, and performing attention calculation and feed-forward neural network processing on the second historical metric data using a second encoder to obtain second intermediate data; processing the first intermediate data using the first attention mechanism to determine first correlation data, and processing the second intermediate data using the second attention mechanism to determine second correlation data.

[0008] In an exemplary embodiment of the present disclosure, performing attention calculation and feed-forward neural network processing on the first historical metric data using the first encoder to obtain first intermediate data, and performing attention calculation and feed-forward neural network processing on the second historical metric data using the second encoder to obtain second intermediate data includes: performing feature transformation on the first historical metric data and the second historical metric data to determine first feature data and second feature data; obtaining a first positional encoding of the first historical metric data and a second positional encoding of the second historical metric data; performing attention calculation and feed-forward neural network processing on the first feature data and the first positional encoding using the first encoder to obtain first intermediate data; performing attention calculation and feed-forward neural network processing on the second feature data and the second positional encoding using the second encoder to obtain second intermediate data.

[0009] In an exemplary embodiment of the present disclosure, the method further includes: processing the first historical metric data using a third attention mechanism to determine third correlation data; processing the first intermediate data using the first attention mechanism to determine first correlation data, and processing the second intermediate data using the second attention mechanism to determine second correlation data includes: processing the first intermediate data and the third correlation data using the first attention mechanism to determine first correlation data; processing the second intermediate data and the third correlation data using the second attention mechanism to determine second correlation data.

[0010] In an exemplary embodiment of the present disclosure, the processing of the first intermediate data and the third correlation data by using the first attention mechanism to determine the first correlation data includes: converting the first intermediate data into first key data and first value data, using the third correlation data as the first query data, and processing the first query data, the first key data, and the first value data by using the first attention mechanism to determine the first correlation data; the processing of the second intermediate data and the third correlation data by using the second attention mechanism to determine the second correlation data includes: converting the second intermediate data into second key data and second value data, using the third correlation data as the second query data, and processing the second query data, the second key data, and the second value data by using the second attention mechanism to determine the second correlation data.

[0011] In an exemplary embodiment of the present disclosure, the relevant indicators include the first relevant indicator of the non-predicted indicator in the target cell and / or the second relevant indicator of other cells associated with the target cell. The second intermediate data includes the third intermediate data corresponding to the first relevant indicator and / or the fourth intermediate data corresponding to the second relevant indicator; the processing of the second intermediate data and the third correlation data by using the second attention mechanism to determine the second correlation data includes: processing the third intermediate data and the third correlation data by using the second attention mechanism to determine the fourth correlation data; and / or processing the fourth intermediate data and the third correlation data by using the second attention mechanism to determine the fifth correlation data; determining the target correlation data according to the first correlation data and the second correlation data includes: determining the target correlation data according to the first correlation data and the fourth correlation data, or the first correlation data and the fifth correlation data, or the first correlation data, the fourth correlation data, and the fifth correlation data.

[0012] In an exemplary embodiment of the present disclosure, after obtaining the first historical indicator data of the to-be-predicted indicator in the target cell and the second historical indicator data of the relevant indicators associated with the target cell, the method further includes: preprocessing the first historical indicator data and the second historical indicator data; performing standardization processing on the preprocessed first historical indicator data and second historical indicator data.

[0013] According to one aspect of the present disclosure, there is provided a communication metric data prediction apparatus, including: a metric data acquisition module configured to acquire first historical metric data of a metric to be predicted in a target cell and second historical metric data of related metrics associated with the target cell; a metric data processing module configured to process the first historical metric data by using a first attention mechanism to determine first correlation data, and process the second historical metric data by using a second attention mechanism to determine second correlation data; a correlation data processing module configured to determine target correlation data according to the first correlation data and the second correlation data; and a metric data prediction module configured to process the target correlation data through a pre-trained communication data prediction network to obtain predicted metric data of the metric to be predicted in the target cell.

[0014] According to one aspect of the present disclosure, there is provided an electronic device, including: a processor; and a memory configured to store executable instructions of the processor; wherein the processor is configured to execute the method according to any one of the above by executing the executable instructions.

[0015] According to one aspect of the present disclosure, there is provided a computer-readable storage medium having a computer program stored thereon, and the computer program, when executed by a processor, implements the method according to any one of the above.

[0016] The exemplary embodiments of the present disclosure have the following beneficial effects:

[0017] In this exemplary embodiment, first historical index data of the index to be predicted in the target cell and second historical index data of related indexes associated with the target cell are obtained; the first historical index data is processed by a first attention mechanism to determine first correlation data, and the second historical index data is processed by a second attention mechanism to determine second correlation data; target correlation data is determined according to the first correlation data and the second correlation data; and the target correlation data is processed by a pre-trained communication data prediction network to obtain predicted index data of the index to be predicted in the target cell. On the one hand, this exemplary embodiment proposes a new method for predicting communication index data. Compared with the prior art in which only historical data of a single index to be predicted can be used to predict index data, this exemplary embodiment can predict the index data of the index to be predicted with more dimensions and richer data based on the first historical index data of the index to be predicted and the second historical index data of related indexes, ensuring the accuracy and effectiveness of index data prediction. On the other hand, this exemplary embodiment respectively uses a first attention mechanism to determine first correlation data and a second attention mechanism to determine second correlation, and determines predicted index data based on the target correlation data. This way of mining index data correlation from different dimensions, fusing the correlation, and reasonably using the target correlation data to predict index data can comprehensively capture the correlation relationship between data, significantly enhancing the accuracy and reliability of index data prediction in wireless communication scenarios.

[0018] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present disclosure and used together with the specification to explain the principles of the present disclosure. Obviously, the following drawings are only some embodiments of the present disclosure, and those of ordinary skill in the art can obtain other drawings without creative efforts based on these drawings.

[0020] Figure 1 Schematically showing a flowchart of a method for predicting communication index data in this exemplary embodiment;

[0021] Figure 2 Schematically showing a framework flow diagram of a method for predicting communication index data in this exemplary embodiment;

[0022] Figure 3 Schematically showing a schematic diagram of a network architecture of a method for predicting communication index data in this exemplary embodiment;

[0023] Figure 4Schematically shows a structural block diagram of a communication metric data prediction device in this exemplary embodiment;

[0024] Figure 5 Schematically shows an electronic device for implementing the above method in this exemplary embodiment. Detailed implementation manners

[0025] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be more complete and comprehensive, and will fully convey the concept of the example embodiments to those skilled in the art. The features, structures, or characteristics described may be combined in any suitable manner in one or more embodiments.

[0026] The exemplary embodiment of the present disclosure first provides a communication metric data prediction method. The application scenario of the method in this embodiment can be: in a wireless communication scenario, predicting metric data such as wireless performance or load for a period of time in the future.

[0027] The following combines the attached Figure 1 For further illustration of this exemplary embodiment, as Figure 1 shown, the communication metric data prediction method may include the following steps S110 to S140:

[0028] Step S110, obtaining first historical metric data of the metric to be predicted in the target cell, and second historical metric data of related metrics associated with the target cell.

[0029] In a communication scenario, a cell refers to an area covered by a base station or a part thereof, and usually refers to a specific geographical area that can be covered by the signal of a base station. Within this area, mobile devices can communicate reliably with the base station through wireless channels. The target cell can be any cell that needs to perform communication metric prediction, which can be determined by a technician or randomly determined by the system. For example, each cell can be taken as the target cell one by one for metric data prediction.

[0030] Metrics can be quantitative metrics for measuring the performance and load status of a communication system, such as throughput, traffic, signal-to-noise ratio, number of users, transmission rate, latency, handover success rate, etc. The metric to be predicted can be any one or more of them. For example, traffic data under the traffic metric can be predicted, or traffic data and the number of users can be predicted, etc.

[0031] The first historical indicator data refers to the historical indicator values of the indicator to be predicted, and the second historical indicator data refers to the historical indicator values of the relevant indicators associated with the target cell. Among them, the relevant indicators associated with the target cell may include other relevant indicators in the target cell except the indicator to be predicted. For example, when the indicator to be predicted is traffic, the relevant indicator may be the number of users. The relevant indicators associated with the target cell may also include the indicators in other cells associated with the target cell. For example, the traffic and the number of users of the adjacent cells of the target cell, etc.

[0032] The historical indicator data can be obtained from the historical records saved in the system, and the specific time range can be customized according to actual needs or user needs. For example, the first historical indicator data and the second historical indicator data within 24 hours from the current time can be obtained, or the first historical indicator data and the second historical indicator data of yesterday and the day before yesterday can be obtained, etc.

[0033] In an exemplary embodiment, after obtaining the first historical indicator data of the indicator to be predicted in the target cell and the second historical indicator data of the relevant indicators associated with the target cell, the above communication indicator data prediction method may further include:

[0034] Preprocess the first historical indicator data and the second historical indicator data;

[0035] Perform standardization processing on the preprocessed first historical indicator data and second historical indicator data.

[0036] To improve the effectiveness and accuracy of data processing, after obtaining the first historical indicator data and the second historical indicator data, they can be preprocessed to make the data reach a complete, accurate, and consistent state. The preprocessing may include cleaning the historical indicator data, such as removing redundant or duplicate data, correcting incorrect data, supplementing or updating missing data, etc., and may also mark outliers or missing values. Then, the preprocessed historical indicator data can be standardized to standardize data with different standards, so that data from different sources and in different formats have consistency and comparability. The specific implementation can be to convert the data into a standard format according to a preset format, or to map the data to a specific interval or satisfy a specific distribution through standardization or normalization methods, etc., for subsequent data analysis and processing.

[0037] Step S120, process the first historical indicator data using the first attention mechanism to determine the first correlation data, and process the second historical indicator data using the second attention mechanism to determine the second correlation data.

[0038] Among them, the first attention mechanism is used to process the first historical metric data from the time dimension. The first correlation data is the result obtained by performing attention calculation in the time dimension, which can reflect the correlation between the first historical metric data in different time series. For example, the first correlation data can be the correlation between the metric data of the same metric in different time series. The second attention mechanism is used to process the second historical metric data from the spatial dimension. The second correlation data is the result obtained by performing attention calculation in the spatial dimension, which can reflect the correlation between the second historical metric data. For example, the to-be-predicted metric of the target cell and other metrics of adjacent cells belong to different spaces, and attention calculation can be performed on the metric data of other metrics of adjacent cells to determine the second correlation data. This exemplary embodiment can determine the second correlation data by calculating the correlation data between other metric data different from the to-be-predicted metric, or the correlation data between other metric data of adjacent cells different from the target cell.

[0039] This exemplary embodiment can configure a first attention module and a second attention module in the network architecture to process through the first attention module and the second attention module to determine the first correlation data and the second correlation data. Among them, the first attention module can be a time cross multi-head attention module, and the second attention module can be a spatial cross multi-head attention module.

[0040] Step S130: Determine the target correlation data according to the first correlation data and the second correlation data.

[0041] Among them, the target correlation data refers to the total correlation data determined after integrating the first correlation data and the second correlation data. This exemplary embodiment can fuse or associate the first correlation data and the second correlation data to obtain the target correlation data. For example, the first correlation data and the second correlation data can be concatenated or integrated into a feature matrix and other methods to obtain the target correlation data.

[0042] Step S140: Process the target correlation data through a pre-trained communication data prediction network to obtain the predicted metric data of the to-be-predicted metric in the target cell.

[0043] Finally, the target correlation data can be processed by a pre-trained communication data prediction network to output the predicted index data of the index to be predicted in the target cell. Among them, the pre-trained communication prediction network can be a convolutional neural network. Taking the target correlation data as input data and inputting it into the communication prediction network can obtain the predicted index data of the index to be predicted. The predicted index data can be the index data of the index to be predicted within a future period predicted by the communication data prediction network according to the input data. For example, in this exemplary embodiment, the index data of the index to be predicted for the next 24 hours can be predicted based on the first historical index data of the index to be predicted and the second historical index data of the related index in the recent 24 hours.

[0044] Based on the above description, in this exemplary embodiment, the first historical index data of the index to be predicted in the target cell and the second historical index data of the related index associated with the target cell are obtained; the first attention mechanism is used to process the first historical index data to determine the first correlation data, and the second attention mechanism is used to process the second historical index data to determine the second correlation data; the target correlation data is determined according to the first correlation data and the second correlation data; the target correlation data is processed by a pre-trained communication data prediction network to obtain the predicted index data of the index to be predicted in the target cell. On the one hand, this exemplary embodiment proposes a new method for predicting communication index data. Compared with the prior art that can only predict index data through the historical data of a single index to be predicted, this exemplary embodiment can predict the index data of the index to be predicted with more dimensions and richer data based on the first historical index data of the index to be predicted and the second historical index data of the related index, ensuring the accuracy and effectiveness of the index data prediction; on the other hand, this exemplary embodiment uses the first attention mechanism to determine the first correlation data and the second attention mechanism to determine the second correlation respectively, and determines the predicted index data based on the target correlation data. This way of mining the correlation of index data from different dimensions, fusing the correlation, and reasonably using the target correlation data to predict the index data can comprehensively capture the correlation relationship between data, significantly enhancing the accuracy and reliability of the index data prediction in the wireless communication scenario.

[0045] In an exemplary embodiment, the above using the first attention mechanism to process the first historical index data to determine the first correlation data and using the second attention mechanism to process the second historical index data to determine the second correlation data may include:

[0046] Using the first encoder to perform attention calculation and feed-forward neural network processing on the first historical index data to obtain the first intermediate data, and using the second encoder to perform attention calculation and feed-forward neural network processing on the second historical index data to obtain the second intermediate data;

[0047] The first attention mechanism is used to process the first intermediate data to determine the first correlation data, and the second attention mechanism is used to process the second intermediate data to determine the second correlation data.

[0048] In this exemplary embodiment, different encoders can be set to perform attention calculation and feed-forward neural network processing on the first historical metric data and the second historical metric data respectively. The structures of the first encoder and the second encoder can be the same. For example, the structure of the encoder can include a multi-head self-attention module, a first residual connection and normalization module, a feed-forward neural network, a second residual connection and normalization module, etc. The structures of the first encoder and the second encoder can also be different, and the specific differences can be set according to actual needs. The present disclosure does not make specific limitations in this regard. Among them, the multi-head self-attention module can be used to perform attention calculation on the first historical metric data, and the feed-forward neural network can be used to perform feed-forward neural network processing on the first historical metric data or the attention calculation result.

[0049] The first intermediate data refers to the output data of the first encoder, and this intermediate data can be the high-dimensional feature data of the first historical metric data. The second intermediate data refers to the output data of the second encoder, and this intermediate data can be the high-dimensional feature data of the second historical metric data.

[0050] In an exemplary embodiment, the above-mentioned use of the first encoder to perform attention calculation and feed-forward neural network processing on the first historical metric data to obtain the first intermediate data, and the use of the second encoder to perform attention calculation and feed-forward neural network processing on the second historical metric data to obtain the second intermediate data may include:

[0051] Perform feature transformation on the first historical metric data and the second historical metric data to determine the first feature data and the second feature data;

[0052] Obtain the first position encoding of the first historical metric data and the second position encoding of the second historical metric data;

[0053] Use the first encoder to perform attention calculation and feed-forward neural network processing on the first feature data and the first position encoding to obtain the first intermediate data;

[0054] Use the second encoder to perform attention calculation and feed-forward neural network processing on the second feature data and the second position encoding to obtain the second intermediate data.

[0055] Among them, feature transformation refers to converting historical indicator data into a feature representation suitable for model processing, which may include data cleaning and feature extraction of historical indicator data, such as word embedding and other processing, to convert the original historical indicator data into a feature vector or feature matrix. For example, mapping the historical indicator data to a low-dimensional vector space, etc. In this exemplary embodiment, feature transformation can be performed on the first historical indicator data to obtain first feature data, and feature transformation can be performed on the second historical indicator data to obtain second feature data.

[0056] In this exemplary embodiment, the encoder can be a Transformer model. Considering that the Transformer model may not be able to perceive the position information of elements in the sequence, position encoding can be introduced to incorporate the position information into the data. Processing the first feature data and the first position encoding using the first encoder may be to first fuse the first feature data with the first position encoding, and then use the first encoder to process the fused feature data to obtain first intermediate data. Processing the second feature data and the second position encoding using the second encoder may be to first fuse the second feature data with the second position encoding, and then use the second encoder to process the fused feature data to obtain second intermediate data. In this exemplary embodiment, the first feature data and the second feature data may be low-dimensional feature data converted from the initial historical indicator data, and the first intermediate data and the second intermediate data may be high-dimensional feature data obtained after being processed by the encoder.

[0057] In an exemplary embodiment, the above communication metric data prediction method further includes:

[0058] Processing the first historical indicator data using a third attention mechanism to determine third correlation data;

[0059] Processing the first intermediate data using a first attention mechanism to determine first correlation data, and processing the second intermediate data using a second attention mechanism to determine second correlation data, including:

[0060] Processing the first intermediate data and the third correlation data using a first attention mechanism to determine first correlation data;

[0061] Processing the second intermediate data and the third correlation data using a second attention mechanism to determine second correlation data.

[0062] In this exemplary embodiment, in addition to using the first attention mechanism to process the first intermediate data processed by the first encoder to determine the first correlation data, the third attention mechanism can also be used to process the first historical indicator data that has not been processed by the encoder to determine the third correlation data, wherein the third correlation data can be the internal correlation of the indicator to be predicted. In this exemplary embodiment, the first historical indicator data can be first subjected to feature transformation, position encoding, and other processing, and then the masked multi-head attention module can be used to perform attention calculation on the processed feature data to obtain the third correlation data. For example, the processed feature data can be first converted into Q (query data), K (key data), and V (value data), and then calculated using the third attention mechanism. Specifically, the intra-indicator correlation, that is, the third correlation data, can be calculated in the following manner:

[0063]

[0064] Among them, Q target ,K target ,V target are the data obtained by converting the feature data obtained by performing feature transformation, position coding and other processing on the first historical indicator data, is a parameter. target ’After that, residual connection and normalization can be performed.

[0065] Furthermore, the first attention mechanism can be used to process the first intermediate data and the third correlation data to determine the first correlation data, and the second attention mechanism can be used to process the second intermediate data and the third correlation data to determine the second correlation data. For example, the first intermediate data can be converted into specific data, combined with the third correlation data, and then the first attention mechanism can be used to process the integrated data to determine the first correlation data, and the same applies to the second correlation data.

[0066] In an exemplary embodiment, the relevant indicator may include a first relevant indicator of the target cell that is not an indicator to be predicted, and / or a second relevant indicator of another cell associated with the target cell, and the second intermediate data may include third intermediate data corresponding to the first relevant indicator, and / or fourth intermediate data corresponding to the second relevant indicator.

[0067] The above-mentioned processing of the second intermediate data and the third correlation data using the second attention mechanism to determine the second correlation data may include:

[0068] Using a second attention mechanism to process the third intermediate data and the third correlation data to determine fourth correlation data; and / or

[0069] Process the fourth intermediate data and the third correlation data using a second attention mechanism to determine the fifth correlation data;

[0070] Further, determining the target correlation data based on the first correlation data and the second correlation data may include:

[0071] Determine the target correlation data according to the first correlation data and the fourth correlation data, or the first correlation data and the fifth correlation data, or the first correlation data, the fourth correlation data, and the fifth correlation data.

[0072] In this exemplary embodiment, the correlation metrics may include other associated metric data different from the metric to be predicted, which may be called different metrics, that is, the first correlation metrics. For example, the target cell may include multiple performance metrics such as traffic, number of users, signal-to-noise ratio, etc. If traffic is used as the metric to be predicted, the number of users and the signal-to-noise ratio may be used as the first correlation metrics.

[0073] The correlation metrics may also include the second correlation metrics of other cells associated with the target cell. Among them, the other cells associated with the target cell may be cells associated with the target cell in terms of region, such as adjacent cells of the target cell, etc. Then, the second correlation metrics may be performance metrics such as traffic, number of users, signal-to-noise ratio, etc. of the adjacent cells, and the second correlation metrics may be called different-cell metrics.

[0074] Further, the second historical metric data may include the second historical metric data corresponding to the different metrics and / or the different-cell metrics. Correspondingly, after encoding the second historical metric data, the intermediate data corresponding to the different metrics, that is, the third intermediate data corresponding to the first correlation metrics, and / or the intermediate data corresponding to the different-cell metrics, that is, the fourth intermediate data corresponding to the second correlation metrics, can be obtained. When using the second attention mechanism to process the second intermediate data and the third correlation data, it may be to process the third intermediate data and the third correlation data respectively to determine the fourth correlation data, and / or to process the fourth intermediate data and the third correlation data to determine the fifth correlation data.

[0075] Finally, when determining the target correlation data, it may be based on the first correlation data and the fourth correlation data, it may be based on the first correlation data and the fifth correlation data, or it may be based on the first correlation data, the fourth correlation data, and the fifth correlation data. The specific method may be determined according to the calculation requirements, the actually obtained data, or the data processing results.

[0076] In an exemplary embodiment, the above-mentioned process of using the first attention mechanism to process the first intermediate data and the third correlation data to determine the first correlation data may include:

[0077] Convert the first intermediate data into first key data and first value data, use the third correlation data as the first query data, and process the first query data, first key data, and first value data using the first attention mechanism to determine the first correlation data;

[0078] The above-mentioned processing of the second intermediate data and the third correlation data using the second attention mechanism to determine the second correlation data may include:

[0079] Convert the second intermediate data into second key data and second value data, use the third correlation data as the second query data, and process the second query data, second key data, and second value data using the second attention mechanism to determine the second correlation data.

[0080] Among them, the first key data and the first value data refer to those obtained by converting the first intermediate data. In this exemplary embodiment, the third correlation data can be used as the first query data, and the first attention mechanism is used to process the first query data, first key data, and first value data to determine the first correlation data. For example, in this exemplary embodiment, the output of the masked multi-head attention module, the third correlation data Q target ′ can be used as the first query data Q, and the correlation between different time series of the same predicted indicator is calculated using the first intermediate data:

[0081]

[0082] Then, perform residual connection and normalization processing.

[0083] Among them, Q target ′ is the output result of the masked multi-head attention module, that is, the third correlation data, K time , V time are the first key data and first value data determined according to the first intermediate data. Different time series of the predicted indicator may refer to sequence data formed by indicator data under the same indicator at different times. For example, the sequence data composed of indicator data of the traffic indicator in the t1-t2 time period and the sequence data composed of indicator data of the traffic indicator in the t3-t4 time period are sequence data of the same indicator at different times.

[0084] The second key data and the second value data refer to those obtained by converting the second intermediate data. In this exemplary embodiment, the second intermediate data can be converted into second key data and second value data, the third correlation data can be used as the second query data, and the second attention mechanism is used to process the second query data, second key data, and second value data to determine the second correlation data.

[0085] If the second intermediate data are respectively the third intermediate data corresponding to the first correlation index and / or the fourth intermediate data corresponding to the second correlation index, the second correlation data can also be separately calculated from these two different types of intermediate data to determine the fourth correlation data and the fifth correlation data.

[0086] For example, in this exemplary embodiment, the output of the masked multi-head attention module, the third correlation data Q target ′ can be used as the second query data Q, and the second correlation data is calculated using the second intermediate data.

[0087] If there are different-indicator data in the target cell that are different from the to-be-predicted indicator, this exemplary embodiment can calculate the fourth correlation between the data of different indicators in the same cell in the following manner:

[0088]

[0089] Then, residual connection and normalization processing are performed.

[0090] Among them, Q target ′ is the output result of the masked multi-head attention module, that is, the third correlation data, and K space , V space refer to the second key data and the second value data determined according to the third intermediate data corresponding to the first correlation index.

[0091] If there are different-cell indicator data of the second correlation index in other cells associated with the target cell, this exemplary embodiment can calculate the fifth correlation data between the different-cell indicator data in the following manner:

[0092]

[0093] Then, residual connection and normalization processing are performed.

[0094] Among them, Q target ′ is the output result of the masked multi-head attention module, that is, the third correlation data, and K cell , V cell refer to the second key data and the second value data determined according to the fourth intermediate data corresponding to the second correlation index.

[0095] When determining the first correlation data Attention time between the same indicator at different time series, the fourth correlation data Attention space between the data of different indicators in the same cell, and the fifth correlation data Attention cellAfter that, the target correlation data can be determined based on the first correlation data and the fourth correlation data, or the first correlation data and the fifth correlation data, or the first correlation data, the fourth correlation data, and the fifth correlation data.

[0096] This exemplary embodiment can perform spatio-temporal attention association on these three types of correlation data. For example, the correlation data in matrix form Attention time ,Attention space ,Attention cell can be concatenated or fused, etc. to obtain the target correlation data. Specifically, the target correlation data can be obtained through the following method:

[0097] Attention total =[Attention time ,Attention space ,Attention cell T

[0098] Furthermore, the feedforward neural network processing, residual connection, and normalization processing, etc. can be performed on the associated target correlation data.

[0099] Finally, the result can be used as input data and input into a pre-trained communication data prediction network to obtain the prediction index data of the index to be predicted.

[0100] Compared with the related technology, the internal connections between the same index at different times and different indexes at the same time cannot be effectively captured. This exemplary embodiment takes into account that when the Transformer model prediction method processes multiple indexes, it is affected by various factors such as signal strength, transmission rate, and network delay. There are complex associations between these factors, and the change trends of the same index at different times are also complex, etc. Therefore, the above method is proposed to calculate from three aspects: the correlation between the same index of the index to be predicted at different time series, the correlation between different indexes in the same cell, and the correlation between different cells. The index data of the index to be predicted is assisted in prediction through the relevant indexes and the indexes of the relevant cells, improving the accuracy of the prediction result and the robustness of the model.

[0101] Figure 2 FIG. shows the framework flowchart of a communication index data prediction method in this exemplary embodiment, which specifically may include:

[0102] ​A data acquisition module 210, configured to acquire data of an index to be predicted, first related index data of the same cell, and second related index data of a different cell associated with the target cell, and perform preprocessing and standardization processing on the data of the index to be predicted, the first related index data, and the second related index data;

[0103] A feature transformation module 220, configured to perform feature transformation on the data of the index to be predicted, the first related index data, and the second related index data;

[0104] A position encoding module 230, configured to perform position encoding on the data of the index to be predicted, the first related index data, and the second related index data;

[0105] A feature fusion module 240, configured to fuse the data after feature transformation and position encoding to obtain feature data corresponding to the data of the index to be predicted, feature data corresponding to the first related index data, and feature data corresponding to the second related index data;

[0106] A first encoder 250, configured to perform attention calculation and feedforward neural network processing on the feature data corresponding to the data of the index to be predicted to obtain first intermediate data;

[0107] A second encoder 260, configured to perform attention calculation and feedforward neural network processing on the feature data corresponding to the first related index data and the feature data corresponding to the second related index data to obtain third intermediate data and fourth intermediate data;

[0108] A decoder 270, configured to process the feature data corresponding to the data of the index to be predicted to obtain third correlation data; process the third intermediate data and the third correlation data to determine fourth correlation data, and / or process the fourth intermediate data and the third correlation data to determine fifth correlation data, and fuse the first correlation data, the fourth correlation data, and / or the fifth correlation data to determine target correlation data;

[0109] A prediction module 280, configured to process the output result of the decoder to predict the predicted index data of the index to be predicted in the target cell.

[0110] Figure 3 The figure shows a schematic diagram of a network architecture of a communication index data prediction method in an exemplary embodiment, which may specifically include the following steps:

[0111] Input the data 302 of the index to be predicted in the target cell into the first feature transformation module 304 to obtain corresponding feature data, and input the first related index data of the same cell and the second related index data 306 of the different cell associated with the target cell into the second feature transformation module 308 to obtain corresponding feature data;

[0112] The result of the first feature transformation module 304 and the result of the first position encoding 310 are input into the first encoder 312 for processing to obtain first intermediate data. The first encoder 312 includes multiple intermediate layers such as multi-head self-attention 3121, residual connection and normalization 3122, feed-forward neural network 3123, and residual connection and normalization 3124;

[0113] The result of the second feature transformation module 308 and the result of the second position encoding 314 are input into the second encoder 316 for processing to obtain second intermediate data. The second intermediate data includes third intermediate data corresponding to the first related index data of the same cell and fourth intermediate data corresponding to the second related index data of the different cell associated with the target cell; The second encoder 316 includes multiple intermediate layers such as multi-head self-attention 3161, residual connection and normalization 3162, feed-forward neural network 3163, and residual connection and normalization 3164;

[0114] The result of the first feature transformation module 304 and the result of the third position encoding 318 are input into the decoder 320 for processing;

[0115] The decoder 320 includes a masked multi-head attention module 3201 and a residual connection and normalization 3202, which are used to perform attention calculation on the result of the first feature transformation module 304 and the result of the first position encoding 318 to determine the third correlation data;

[0116] The decoder 320 includes a temporal cross multi-head attention module 3203 and a residual connection and normalization 3204, which are used to perform attention calculation on the first intermediate data and the third correlation data to determine the first correlation data;

[0117] The decoder 320 includes a spatial cross multi-head attention module 3205 and a residual connection and normalization 3206, which are used to perform attention calculation on the second intermediate data and the third correlation data to determine the second correlation data; When the second intermediate data includes the third intermediate data and / or the fourth intermediate data, it may be to process the third intermediate data and the third correlation data to determine the fourth correlation data; and / or process the fourth intermediate data and the third correlation data to determine the fifth correlation data;

[0118] The decoder 320 includes a spatio-temporal attention association module 3207, a feed-forward neural network 3208, and a residual connection and normalization process 3209, which are used to associate the first correlation data and the second correlation data to obtain and output the target correlation data. Specifically, it may be to associate the first correlation data, the fourth correlation data, and the fifth correlation data to obtain and output the target correlation data;

[0119] A convolutional neural network 322 for processing target correlation data;

[0120] A prediction output module 324 for outputting prediction index data of an index to be predicted.

[0121] Based on the above method, on the one hand, in this exemplary embodiment, through the first encoder and the second encoder, different indexes in the same cell and indexes of adjacent cells can be independently encoded, and an association operation can be implemented in the cross-attention mechanism of the decoder, effectively integrating the intra-index correlation and the inter-index correlation, and significantly improving the prediction accuracy; in addition, the weighted output of the intra-index and inter-index correlations can be synthesized by using a convolutional neural network, combined with the index information of adjacent cells, to assist in inferring the index to be predicted. When there are missing values or outliers in the target index, the model can perform compensatory prediction through relevant indexes and data of neighboring cells, thereby improving the accuracy and robustness of the prediction result.

[0122] On the other hand, compared with the traditional Transformer model that captures dependencies in time series through the self-attention mechanism, the cross-attention mechanism of the second encoder for different indexes / different cells can solve the problem that the traditional Transformer model is difficult to capture complex multi-dimensional connections, and can more comprehensively model the spatio-temporal dependencies in the wireless communication network; in addition, compared with traditional methods (such as LSTM, GRU or the classic Transformer) that lead to unstable prediction results and large errors when facing missing values or noises common in wireless communication data, in this exemplary embodiment, by integrating the correlations of different indexes in the same cell and the auxiliary information of adjacent cells, compensatory inference can be performed based on multi-dimensional association information, significantly enhancing the adaptability of the model in complex wireless communication scenarios, and having stronger practicability and reliability.

[0123] The exemplary embodiment of the present disclosure also provides a communication index data prediction device. Referring to Figure 4 , the device 400 may include an index data acquisition module 410 for acquiring first historical index data of an index to be predicted in a target cell and second historical index data of relevant indexes associated with the target cell; an index data processing module 420 for processing the first historical index data by using a first attention mechanism to determine first correlation data, and processing the second historical index data by using a second attention mechanism to determine second correlation data; a correlation data processing module 430 for determining target correlation data according to the first correlation data and the second correlation data; and an index data prediction module 440 for processing the target correlation data through a pre-trained communication data prediction network to obtain prediction index data of the index to be predicted in the target cell.

[0124] In an exemplary embodiment, the metric data processing module 420 includes: an intermediate data determination unit configured to perform attention calculation and feedforward neural network processing on first historical metric data using a first encoder to obtain first intermediate data, and perform attention calculation and feedforward neural network processing on second historical metric data using a second encoder to obtain second intermediate data; a correlation data determination unit configured to process the first intermediate data using a first attention mechanism to determine first correlation data, and process the second intermediate data using a second attention mechanism to determine second correlation data.

[0125] In an exemplary embodiment, the intermediate data determination unit includes: a feature transformation subunit configured to perform feature transformation on the first historical metric data and the second historical metric data to determine first feature data and second feature data; a position encoding acquisition subunit configured to acquire a first position encoding of the first historical metric data and a second position encoding of the second historical metric data; a first calculation subunit configured to perform attention calculation and feedforward neural network processing on the first feature data and the first position encoding using a first encoder to obtain first intermediate data; a second calculation subunit configured to perform attention calculation and feedforward neural network processing on the second feature data and the second position encoding using a second encoder to obtain second intermediate data.

[0126] In an exemplary embodiment, the communication metric data prediction device further includes: a third calculation unit configured to process the first historical metric data using a third attention mechanism to determine third correlation data; the correlation data determination unit includes: a first correlation determination subunit configured to process the first intermediate data and the third correlation data using a first attention mechanism to determine first correlation data; a second correlation determination subunit configured to process the second intermediate data and the third correlation data using a second attention mechanism to determine second correlation data.

[0127] In an exemplary embodiment, the first correlation determination subunit includes: a first data calculation subunit configured to convert the first intermediate data into first key data and first value data, use the third correlation data as first query data, and process the first query data, the first key data, and the first value data using a first attention mechanism to determine first correlation data; the second correlation determination subunit includes: a second data calculation subunit configured to convert the second intermediate data into second key data and second value data, use the third correlation data as second query data, and process the second query data, the second key data, and the second value data using a second attention mechanism to determine second correlation data.

[0128] In an exemplary embodiment, the relevant metrics include first relevant metrics of non-predicted metrics in the target cell, and / or second relevant metrics of other cells associated with the target cell. The second intermediate data includes third intermediate data corresponding to the first relevant metrics, and / or fourth intermediate data corresponding to the second relevant metrics. The second correlation determination subunit includes: an other correlation data determination subunit, configured to process the third intermediate data and third correlation data by using a second attention mechanism to determine fourth correlation data; and / or process the fourth intermediate data and third correlation data by using a second attention mechanism to determine fifth correlation data. The correlation data processing module 430 includes: a correlation data association unit, configured to determine target correlation data according to the first correlation data and the fourth correlation data, or the first correlation data and the fifth correlation data, or the first correlation data, the fourth correlation data, and the fifth correlation data.

[0129] In an exemplary embodiment, after obtaining first historical metric data of a metric to be predicted in a target cell and second historical metric data of relevant metrics associated with the target cell, the communication metric data prediction device further includes: a preprocessing unit, configured to preprocess the first historical metric data and the second historical metric data; and a normalization unit, configured to perform normalization processing on the preprocessed first historical metric data and second historical metric data.

[0130] The specific details of each module / unit in the above device have been described in detail in the embodiments of the method part. The details not disclosed can be referred to the embodiments of the method part, so they will not be elaborated here.

[0131] An exemplary embodiment of the present disclosure further provides an electronic device capable of implementing the above method.

[0132] Those skilled in the art can understand that various aspects of the present disclosure can be implemented as a system, a method, or a program product. Therefore, various aspects of the present disclosure can be specifically implemented in the following forms, namely: a complete hardware implementation manner, a complete software implementation manner (including firmware, microcode, etc.), or an implementation manner combining hardware and software aspects, which can be collectively referred to as "circuit", "module", or "system" here.

[0133] Next, refer to Figure 5 to describe the electronic device 500 according to this exemplary embodiment of the present disclosure. Figure 5 The displayed electronic device 500 is only an example and should not impose any limitations on the functions and usage scope of the embodiments of the present disclosure.

[0134] As Figure 5As shown, the electronic device 500 is presented in the form of a general-purpose computing device. The components of the electronic device 500 may include, but are not limited to: at least one of the above-mentioned processing units 510, at least one of the above-mentioned storage units 520, a bus 530 connecting different system components (including the storage unit 520 and the processing unit 510), and a display unit 540.

[0135] Among them, the storage unit stores program code, and the program code can be executed by the processing unit 510, so that the processing unit 510 executes the steps according to various exemplary embodiments of the present disclosure described in the "Exemplary Method" section of the present specification above. For example, the processing unit 510 can execute Figure 1 the steps shown, etc.

[0136] The storage unit 520 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 521 and / or a cache storage unit 522, and may further include a read-only storage unit (ROM) 523.

[0137] The storage unit 520 may further include a program / utilities 524 having a set (at least one) of program modules 525. Such program modules 525 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment.

[0138] The bus 530 may represent one or more of several types of bus structures, including a storage unit bus or a storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any bus structure in a variety of bus structures.

[0139] The electronic device 500 can also communicate with one or more external devices 600 (such as a keyboard, a pointing device, a Bluetooth device, etc.), can also communicate with one or more devices that enable a user to interact with the electronic device 500, and / or communicate with any device that enables the electronic device 500 to communicate with one or more other computing devices (such as a router, a modem, etc.). Such communication can be carried out through an input / output (I / O) interface 550. And, the electronic device 500 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 560. As shown in the figure, the network adapter 560 communicates with other modules of the electronic device 500 through the bus 530. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in combination with the electronic device 500, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0140] Through the description of the above embodiments, those skilled in the art can easily understand that the exemplary embodiments described herein can be implemented by software or by a combination of software and necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the exemplary embodiments of the present disclosure.

[0141] The exemplary embodiments of the present disclosure also provide a computer-readable storage medium, on which a program product capable of implementing the above methods of this specification is stored. In some possible embodiments, various aspects of the present disclosure can also be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to enable the terminal device to execute the steps according to the various exemplary embodiments described in the above "Exemplary Methods" section of this specification.

[0142] The exemplary embodiments of the present disclosure also provide a program product for implementing the above methods. It can be a portable compact disc read-only memory (CD-ROM) and includes program code, and can run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited to this. In this document, the readable storage medium can be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, device, or device.

[0143] The program product can adopt any combination of one or more readable media. The readable media can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0144] A computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which readable program code is carried. Such a propagated data signal may take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing. The readable signal medium may also be any readable medium other than a readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0145] The program code contained on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0146] The program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and also including conventional procedural programming languages such as the "C" language or similar programming languages. The program code may execute entirely on the user's computing device, partially on the user's device, execute as a stand-alone software package, execute partially on the user's computing device and partially on a remote computing device, or execute entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0147] Furthermore, the above-mentioned drawings are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present disclosure, and are not for limiting purposes. It is easily understood that the processes shown in the above-mentioned drawings do not indicate or limit the chronological order of these processes. Additionally, it is also easily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0148] It should be noted that although several modules or units of a device for action execution are mentioned in the above detailed description, such a division is not mandatory. In fact, according to the exemplary embodiments of the present disclosure, the features and functions of two or more of the above-mentioned modules or units may be embodied in one module or unit. Conversely, the features and functions of one module or unit described above may be further divided and embodied by multiple modules or units.

[0149] Other embodiments of the present disclosure will be readily apparent to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include known common general knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and examples are only to be considered as exemplary, and the true scope and spirit of the present disclosure are pointed out by the claims.

[0150] It should be understood that the present disclosure is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only defined by the appended claims.

Claims

1. A method for predicting communication metric data, characterized in that Including: Obtain the first historical index data of the target index in the target cell, and the second historical index data of the relevant index associated with the target cell; Process the first historical index data by using a first attention mechanism to determine first correlation data, and process the second historical index data by using a second attention mechanism to determine second correlation data; Determine target correlation data according to the first correlation data and the second correlation data; Process the target correlation data through a pre-trained communication data prediction network to obtain the predicted index data of the target index in the target cell.

2. The method according to claim 1, wherein The step of processing the first historical index data by using a first attention mechanism to determine first correlation data, and processing the second historical index data by using a second attention mechanism to determine second correlation data includes: Perform attention calculation and feed-forward neural network processing on the first historical index data by using a first encoder to obtain first intermediate data, and perform attention calculation and feed-forward neural network processing on the second historical index data by using a second encoder to obtain second intermediate data; Process the first intermediate data by using a first attention mechanism to determine first correlation data, and process the second intermediate data by using a second attention mechanism to determine second correlation data.

3. The method according to claim 2, wherein The step of performing attention calculation and feed-forward neural network processing on the first historical index data by using a first encoder to obtain first intermediate data, and performing attention calculation and feed-forward neural network processing on the second historical index data by using a second encoder to obtain second intermediate data includes: Perform feature transformation on the first historical index data and the second historical index data to determine first feature data and second feature data; Obtain the first position encoding of the first historical index data and the second position encoding of the second historical index data; Perform attention calculation and feed-forward neural network processing on the first feature data and the first position encoding by using a first encoder to obtain first intermediate data; Perform attention calculation and feed-forward neural network processing on the second feature data and the second position encoding by using a second encoder to obtain second intermediate data.

4. The method according to claim 2, characterized in that, The method further includes: Process the first historical index data by using a third attention mechanism to determine third correlation data; The step of processing the first intermediate data by using a first attention mechanism to determine first correlation data, and processing the second intermediate data by using a second attention mechanism to determine second correlation data includes: Process the first intermediate data and the third correlation data by using a first attention mechanism to determine first correlation data; Process the second intermediate data and the third correlation data by using a second attention mechanism to determine second correlation data.

5. The method according to claim 4, characterized in that, The step of processing the first intermediate data and the third correlation data by using a first attention mechanism to determine first correlation data includes: Convert the first intermediate data into first key data and first value data, use the third correlation data as the first query data, and process the first query data, first key data, and first value data using a first attention mechanism to determine the first correlation data; The processing the second intermediate data and the third correlation data using a second attention mechanism to determine the second correlation data includes: Convert the second intermediate data into second key data and second value data, use the third correlation data as the second query data, and process the second query data, second key data, and second value data using a second attention mechanism to determine the second correlation data.

6. The method according to claim 4, wherein The correlation metrics include a first correlation metric of non-predicted metrics in the target cell and / or a second correlation metric of other cells associated with the target cell. The second intermediate data includes third intermediate data corresponding to the first correlation metric and / or fourth intermediate data corresponding to the second correlation metric; The processing the second intermediate data and the third correlation data using a second attention mechanism to determine the second correlation data includes: Process the third intermediate data and the third correlation data using a second attention mechanism to determine the fourth correlation data; and / or Process the fourth intermediate data and the third correlation data using a second attention mechanism to determine the fifth correlation data; The determining the target correlation data according to the first correlation data and the second correlation data includes: Determine the target correlation data according to the first correlation data and the fourth correlation data, or the first correlation data and the fifth correlation data, or the first correlation data, the fourth correlation data, and the fifth correlation data.

7. The method according to claim 1, characterized in that, After obtaining the first historical metric data of the metric to be predicted in the target cell and the second historical metric data of the correlation metrics associated with the target cell, the method further includes: Preprocess the first historical metric data and the second historical metric data; Perform normalization processing on the preprocessed first historical metric data and second historical metric data.

8. A communication metric data prediction device, characterized in that, Including: A metric data acquisition module, configured to acquire the first historical metric data of the metric to be predicted in the target cell and the second historical metric data of the correlation metrics associated with the target cell; A metric data processing module, configured to process the first historical metric data using a first attention mechanism to determine the first correlation data, and process the second historical metric data using a second attention mechanism to determine the second correlation data; A correlation data processing module, configured to determine the target correlation data according to the first correlation data and the second correlation data; A metric data prediction module, configured to process the target correlation data through a pre-trained communication data prediction network to obtain the predicted metric data of the metric to be predicted in the target cell.

9. An electronic device, characterized in that, Including: A processor; And A memory, configured to store executable instructions of the processor; Wherein, the processor is configured to execute the method according to any one of claims 1-7 by executing the executable instructions.

10. A computer-readable storage medium, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method according to any one of claims 1-7.

Images