Intelligent identification methods, electronic devices, apparatuses and storage media for high-load cells

Abstract

This application provides a method, electronic device, apparatus, and storage medium for intelligent identification of high-load cells. The method includes: acquiring time-series data of multiple load characteristic indicators of a cell to be identified within a historical time period; inputting the time-series data into an Informer model to obtain a multi-indicator prediction result output by the Informer model, wherein the multi-indicator prediction result includes the prediction results of the multiple load characteristic indicators of the cell to be identified within the prediction time period; and determining whether the cell to be identified is a high-load cell based on the prediction results and a preset high-load standard. The Informer model is trained based on historical time-series data of multiple sample cells, and the historical time-series data of the multiple load characteristic indicators includes historical time-series data of the multiple load characteristic indicators. This enables intelligent identification of massive numbers of cells, significantly improving the efficiency and accuracy of high-load cell identification.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, and in particular to a method, electronic device, apparatus and storage medium for intelligent identification of high-load cells. Background Technology

[0002] The explosive growth of data traffic has put enormous pressure on wireless networks, directly impacting user experience. How to ensure a good user experience while reducing investment costs and improving expansion efficiency has become an urgent issue for network optimization. This largely depends on the identification of high-load cells. If high-load cells can be intelligently identified, then timely optimization and expansion can be implemented. Summary of the Invention

[0003] This application provides a method, electronic device, apparatus, and storage medium for intelligent identification of high-load cells, which can be used to intelligently identify high-load cells and improve capacity expansion efficiency.

[0004] Firstly, this application provides a method for intelligent identification of high-load cells, including:

[0005] Obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical period;

[0006] Input the time series data into the Informer model to obtain the multi-indicator prediction results output by the Informer model. The multi-indicator prediction results include the prediction results of multiple load characteristic indicators of the cell to be identified within the prediction time period.

[0007] Based on the prediction results and the preset high load criteria, determine whether the cell to be identified is a high load cell;

[0008] The Informer model is trained based on multi-indicator historical time-series data from multiple sample cells. The multi-indicator historical time-series data includes historical time-series data of multiple load characteristic indicators.

[0009] In some embodiments, obtaining time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period includes:

[0010] The original data of multiple load characteristic indicators of the cell to be identified within a historical time period are cleaned to obtain the first processed data of multiple load characteristic indicators.

[0011] Based on the first processed data, it is determined whether the cell to be identified is an abnormal cell; abnormal cells include cells with a missing sampling point ratio greater than the first threshold, out-of-service cells, and broadcast test cells;

[0012] If it is determined that the cell to be identified is not an abnormal cell, the abnormal sampling points and missing sampling points in the first processed data are numerically processed to obtain the second processed data with multiple load characteristic indicators.

[0013] The second processing data is standardized to obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical period.

[0014] In some embodiments, numerical processing is performed on outlier and missing sample points in the first processed data, including:

[0015] Based on the Prophet model, the predicted data of various load characteristic indicators of the cell to be identified in the historical time period are obtained respectively.

[0016] Based on the predicted data and the preset confidence interval, identify the abnormal sampling points in the first processed data;

[0017] The predicted data is used to fill in the outlier and missing sample points in the first processed data.

[0018] In some embodiments, the method for determining a cell that has gone out of service includes:

[0019] If a community has a monthly average traffic value of 0 for at least one month, then the community is considered to be out of service.

[0020] In some embodiments, the method for determining the broadcast test cell includes:

[0021] If the ratio between the maximum monthly average traffic value and the minimum monthly average traffic value of a cell is greater than the second threshold, then the cell is a broadcast test cell.

[0022] In some embodiments, load characteristics include uplink traffic of the cell, downlink traffic of the cell, average uplink physical resource block (PRB) utilization of the cell, average downlink PRB utilization of the cell, and physical downlink control channel (PDCCH) control channel element (CCE) occupancy rate of the cell.

[0023] In some embodiments, the high load standard includes an uplink high load standard and a downlink high load standard;

[0024] The uplink high load criteria include: if the average uplink traffic of the cell is greater than the uplink traffic threshold and the average uplink PRB utilization rate of the cell is greater than the average uplink PRB utilization rate threshold during the forecast period, then the cell meets the uplink high load criteria.

[0025] The downlink high load criteria include: during the prediction period, the average downlink traffic of the cell is greater than the downlink traffic threshold, and at least one of the following is met: the average downlink PRB utilization rate of the cell is greater than the downlink PRB average utilization rate threshold, or the cell's PDCCH channel CCE occupancy rate is greater than the PDCCH channel CCE occupancy rate threshold.

[0026] In some embodiments, determining whether a cell to be identified is a high-load cell based on prediction results and a preset high-load standard includes:

[0027] If the prediction results determine that the cell to be identified meets the uplink high load standard and / or downlink high load standard, then the cell to be identified is determined to be a high load cell.

[0028] In some embodiments, the encoder of the Informer model includes a multi-layer coding sub-module, each of which includes a multi-head probabilistic sparse self-attention sub-layer and a self-attention distillation sub-layer.

[0029] The Informer model's decoder consists of multiple decoding sub-modules, each of which includes a masked multi-head probabilistic sparse self-attention sub-layer and a multi-head attention sub-layer.

[0030] The encoder's output features are input into the decoder's mask multi-head probability sparse self-attention sublayer.

[0031] In some embodiments, the training loss function of the Informer model includes the Log-Cosh loss function.

[0032] In some embodiments, the encoder of the Informer model includes a 3-layer encoding submodule, and the decoder of the Informer model includes a 2-layer decoding submodule.

[0033] Secondly, this application also provides an electronic device, including a memory, a transceiver, and a processor;

[0034] A memory for storing computer programs; a transceiver for sending and receiving data under the control of the processor; and a processor for reading the computer programs from the memory and performing the following operations:

[0035] Obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period;

[0036] Input the time series data into the Informer model to obtain the multi-indicator prediction results output by the Informer model. The multi-indicator prediction results include the prediction results of multiple load characteristic indicators of the cell to be identified within the prediction time period.

[0037] Based on the prediction results and the preset high load criteria, determine whether the cell to be identified is a high load cell;

[0038] The Informer model is trained based on multi-indicator historical time-series data from multiple sample cells. The multi-indicator historical time-series data includes historical time-series data of multiple load characteristic indicators.

[0039] In some embodiments, obtaining time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period includes:

[0040] The original data of multiple load characteristic indicators of the cell to be identified within a historical time period are cleaned to obtain the first processed data of multiple load characteristic indicators.

[0041] Based on the first processed data, it is determined whether the cell to be identified is an abnormal cell; abnormal cells include cells with a missing sampling point ratio greater than the first threshold, out-of-service cells, and broadcast test cells;

[0042] If it is determined that the cell to be identified is not an abnormal cell, the abnormal sampling points and missing sampling points in the first processed data are numerically processed to obtain the second processed data with multiple load characteristic indicators.

[0043] The second processing data is standardized to obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical period.

[0044] In some embodiments, numerical processing is performed on outlier and missing sample points in the first processed data, including:

[0045] Based on the Prophet model, the predicted data of various load characteristic indicators of the cell to be identified in the historical time period are obtained respectively.

[0046] Based on the predicted data and the preset confidence interval, identify the abnormal sampling points in the first processed data;

[0047] The predicted data is used to fill in the outlier and missing sample points in the first processed data.

[0048] In some embodiments, the method for determining a cell that has gone out of service includes:

[0049] If a community has a monthly average traffic value of 0 for at least one month, then the community is considered to be out of service.

[0050] In some embodiments, the method for determining the broadcast test cell includes:

[0051] If the ratio between the maximum monthly average traffic value and the minimum monthly average traffic value of a cell is greater than the second threshold, then the cell is a broadcast test cell.

[0052] In some embodiments, load characteristics include uplink traffic of the cell, downlink traffic of the cell, average uplink physical resource block (PRB) utilization of the cell, average downlink PRB utilization of the cell, and physical downlink control channel (PDCCH) control channel element (CCE) occupancy rate of the cell.

[0053] In some embodiments, the high load standard includes an uplink high load standard and a downlink high load standard;

[0054] The uplink high load criteria include: if the average uplink traffic of the cell is greater than the uplink traffic threshold and the average uplink PRB utilization rate of the cell is greater than the average uplink PRB utilization rate threshold during the forecast period, then the cell meets the uplink high load criteria.

[0055] The downlink high load criteria include: during the prediction period, the average downlink traffic of the cell is greater than the downlink traffic threshold, and at least one of the following is met: the average downlink PRB utilization rate of the cell is greater than the downlink PRB average utilization rate threshold, or the cell's PDCCH channel CCE occupancy rate is greater than the PDCCH channel CCE occupancy rate threshold.

[0056] In some embodiments, determining whether a cell to be identified is a high-load cell based on prediction results and a preset high-load standard includes:

[0057] If the prediction results determine that the cell to be identified meets the uplink high load standard and / or downlink high load standard, then the cell to be identified is determined to be a high load cell.

[0058] In some embodiments, the encoder of the Informer model includes a multi-layer coding sub-module, each of which includes a multi-head probabilistic sparse self-attention sub-layer and a self-attention distillation sub-layer.

[0059] The Informer model's decoder consists of multiple decoding sub-modules, each of which includes a masked multi-head probabilistic sparse self-attention sub-layer and a multi-head attention sub-layer.

[0060] The encoder's output features are input into the decoder's mask multi-head probability sparse self-attention sublayer.

[0061] In some embodiments, the training loss function of the Informer model includes the Log-Cosh loss function.

[0062] In some embodiments, the encoder of the Informer model includes a 3-layer encoding submodule, and the decoder of the Informer model includes a 2-layer decoding submodule.

[0063] Thirdly, this application also provides a high-load cell intelligent identification device, comprising:

[0064] The acquisition unit is used to acquire time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period;

[0065] The prediction unit is used to input time series data into the Informer model and obtain the multi-indicator prediction results output by the Informer model. The multi-indicator prediction results include the prediction results of multiple load characteristic indicators of the cell to be identified within the prediction time period.

[0066] The determination unit is used to determine whether the cell to be identified is a high-load cell based on the prediction results and the preset high-load criteria.

[0067] The Informer model is trained based on multi-indicator historical time-series data from multiple sample cells. The multi-indicator historical time-series data includes historical time-series data of multiple load characteristic indicators.

[0068] In some embodiments, obtaining time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period includes:

[0069] The original data of multiple load characteristic indicators of the cell to be identified within a historical time period are cleaned to obtain the first processed data of multiple load characteristic indicators.

[0070] Based on the first processed data, it is determined whether the cell to be identified is an abnormal cell; abnormal cells include cells with a missing sampling point ratio greater than the first threshold, out-of-service cells, and broadcast test cells;

[0071] If it is determined that the cell to be identified is not an abnormal cell, the abnormal sampling points and missing sampling points in the first processed data are numerically processed to obtain the second processed data with multiple load characteristic indicators.

[0072] The second processing data is standardized to obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical period.

[0073] In some embodiments, numerical processing is performed on outlier and missing sample points in the first processed data, including:

[0074] Based on the Prophet model, the predicted data of various load characteristic indicators of the cell to be identified in the historical time period are obtained respectively.

[0075] Based on the predicted data and the preset confidence interval, identify the abnormal sampling points in the first processed data;

[0076] The predicted data is used to fill in the outlier and missing sample points in the first processed data.

[0077] In some embodiments, the method for determining a cell that has gone out of service includes:

[0078] If a community has a monthly average traffic value of 0 for at least one month, then the community is considered to be out of service.

[0079] In some embodiments, the method for determining the broadcast test cell includes:

[0080] If the ratio between the maximum monthly average traffic value and the minimum monthly average traffic value of a cell is greater than the second threshold, then the cell is a broadcast test cell.

[0081] In some embodiments, load characteristics include uplink traffic of the cell, downlink traffic of the cell, average uplink physical resource block (PRB) utilization of the cell, average downlink PRB utilization of the cell, and physical downlink control channel (PDCCH) control channel element (CCE) occupancy rate of the cell.

[0082] In some embodiments, the high load standard includes an uplink high load standard and a downlink high load standard;

[0083] The uplink high load criteria include: if the average uplink traffic of the cell is greater than the uplink traffic threshold and the average uplink PRB utilization rate of the cell is greater than the average uplink PRB utilization rate threshold during the forecast period, then the cell meets the uplink high load criteria.

[0084] The downlink high load criteria include: during the prediction period, the average downlink traffic of the cell is greater than the downlink traffic threshold, and at least one of the following is met: the average downlink PRB utilization rate of the cell is greater than the downlink PRB average utilization rate threshold, or the cell's PDCCH channel CCE occupancy rate is greater than the PDCCH channel CCE occupancy rate threshold.

[0085] In some embodiments, determining whether a cell to be identified is a high-load cell based on prediction results and a preset high-load standard includes:

[0086] If the prediction results determine that the cell to be identified meets the uplink high load standard and / or downlink high load standard, then the cell to be identified is determined to be a high load cell.

[0087] In some embodiments, the encoder of the Informer model includes a multi-layer coding sub-module, each of which includes a multi-head probabilistic sparse self-attention sub-layer and a self-attention distillation sub-layer.

[0088] The Informer model's decoder consists of multiple decoding sub-modules, each of which includes a masked multi-head probabilistic sparse self-attention sub-layer and a multi-head attention sub-layer.

[0089] The encoder's output features are input into the decoder's mask multi-head probability sparse self-attention sublayer.

[0090] In some embodiments, the training loss function of the Informer model includes the Log-Cosh loss function.

[0091] In some embodiments, the encoder of the Informer model includes a 3-layer encoding submodule, and the decoder of the Informer model includes a 2-layer decoding submodule.

[0092] Fourthly, this application also provides a non-transiently readable storage medium storing a computer program for causing a processor to execute the high-load cell intelligent identification method described in the first aspect above.

[0093] Fifthly, this application also provides a communication device, wherein the communication device stores a computer program, the computer program being used to cause the communication device to perform the high-load cell intelligent identification method described in the first aspect above.

[0094] In a sixth aspect, this application also provides a processor-readable storage medium storing a computer program for causing a processor to execute the high-load cell intelligent identification method described in the first aspect above.

[0095] In a seventh aspect, this application also provides a chip product storing a computer program for causing the chip product to execute the high-load cell intelligent identification method described in the first aspect above.

[0096] The high-load cell intelligent identification method, electronic device, apparatus and storage medium provided in this application use the Informer model to predict multiple load characteristic indicators of the cell to be identified, and then determine the high-load cell based on the prediction results and the preset high-load standard. It can realize the intelligent identification of a large number of cells without the need for a large investment of human resources, which greatly improves the efficiency and accuracy of high-load cell identification. Attached Figure Description

[0097] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0098] Figure 1 A flowchart illustrating the intelligent identification method for high-load cells provided in this application embodiment;

[0099] Figure 2Example diagram of daily total traffic index data for out-of-service cells and broadcast testing cells provided in the embodiments of this application;

[0100] Figure 3 Example diagram of the results of downlink traffic prediction using the Prophet model provided in the embodiments of this application;

[0101] Figure 4 Comparison diagram of downlink traffic abnormal sampling points and missing sampling points before and after processing provided in the embodiments of this application;

[0102] Figure 5 A flowchart of Informer model training and prediction provided for embodiments of this application;

[0103] Figure 6 The diagram shows the Informer model structure and training example provided in the embodiments of this application;

[0104] Figure 7 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application;

[0105] Figure 8 This is a schematic diagram of the structure of the intelligent identification device for high-load cells provided in the embodiments of this application. Detailed Implementation

[0106] In the embodiments of this application, the term "and / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following associated objects have an "or" relationship.

[0107] In the embodiments of this application, the term "multiple" refers to two or more, and other quantifiers are similar.

[0108] In the embodiments of this application, the terms "first," "second," etc., are used to distinguish similar objects, and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, and the number of objects is not limited; for example, the first object can be one or more.

[0109] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0110] Identifying high-load cells manually requires significant manpower, is time-consuming, and cumbersome, thus limiting its applicability to small-scale operations and large-scale network-wide optimization and expansion. Collecting cell traffic volume information and using artificial intelligence (AI) models to predict future traffic is a feasible approach. Commonly used AI models include Auto-Regressive Moving Average (ARMA), Prophet, and Long Short-Term Memory (LSTM) time series forecasting models. However, these models have limitations in predicting high-load cells. For example, ARMA and Prophet models can only consider single elements for single-indicator prediction, while communication indicators are interdependent and interconnected; single-indicator time prediction cannot meet application requirements. While LSTM models can consider multiple features, their performance in predicting long periods is generally poor; for instance, predicting related indicators for the next month using LSTM models yields relatively low accuracy.

[0111] To address the aforementioned issues, the embodiments of this application provide a solution that uses the Informer model (or an improved Informer model) to predict the long and short cycles of multiple load characteristic indicators of a cell. Based on the prediction results and a preset high-load standard, high-load cells are intelligently identified, which can effectively improve capacity expansion efficiency and ensure user experience.

[0112] Figure 1 This is a flowchart illustrating the intelligent identification method for high-load cells provided in the embodiments of this application, as shown below. Figure 1 As shown, the method includes the following steps:

[0113] Step 100: Obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period.

[0114] Specifically, the execution entity of each step in this method can be a high-load cell intelligent identification device, which can be implemented by software and / or hardware. The device can be integrated into an electronic device, which can be a terminal device (such as a smartphone, personal computer, etc.), a server (such as a local server or cloud server, or a server cluster, etc.), a network device (e.g., a base station) or other devices.

[0115] When identifying high-load cells, for the cell to be identified, time-series data of multiple load characteristic indicators for that cell over a historical period are first obtained. The historical period can be selected according to the Informer model's requirements for input data and is not limited here; for example, it could be data from the past year. Load characteristic indicators refer to features that can be used to characterize the cell's load. The load characteristic indicators used for Informer model training and inference are consistent. Time-series data, also known as time-series data, is data arranged in chronological order. For example, the daily data of a certain load characteristic indicator for a cell over the past year, arranged in chronological order, constitutes the time-series data for that load characteristic indicator for that cell over the past year.

[0116] In some embodiments, load characteristics include uplink traffic, downlink traffic, average uplink physical resource block (PRB) utilization, average downlink PRB utilization, and physical downlink control channel (PDCCH) control channel element (CCE) occupancy. Data for these load characteristics can be obtained by collecting key performance indicator (KPI) data for the cell.

[0117] In some embodiments, obtaining time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period includes:

[0118] The original data of multiple load characteristic indicators of the cell to be identified within a historical time period are cleaned to obtain the first processed data of multiple load characteristic indicators.

[0119] Based on the first processed data, it is determined whether the cell to be identified is an abnormal cell; abnormal cells include cells with a missing sampling point ratio greater than the first threshold, out-of-service cells, and broadcast test cells;

[0120] If it is determined that the cell to be identified is not an abnormal cell, the abnormal sampling points and missing sampling points in the first processed data are numerically processed to obtain the second processed data with multiple load characteristic indicators.

[0121] The second processing data is standardized to obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical period.

[0122] Specifically, since the collected raw data may contain issues such as time-series disorder, missing timestamps, abnormal sampling points, and missing sampling points, it is necessary to preprocess the raw data of each load characteristic indicator. This includes data cleaning, abnormal cell removal, and handling of abnormal and missing sampling points. The time-series data of each load characteristic indicator obtained after preprocessing are then input into the Informer model for multi-indicator prediction. The training dataset used to train the Informer model is also obtained through the same preprocessing procedure.

[0123] Data cleaning refers to the preliminary processing of the raw data for each load characteristic indicator, including sorting the data in chronological order and filling in missing timestamps (the indicator value corresponding to the missing timestamp is empty), etc.

[0124] After data cleaning, the next step is to determine whether the cell to be identified is an abnormal cell based on the first processed data of multiple load characteristic indicators obtained after data cleaning. If it is an abnormal cell, no prediction will be made for that cell. During training, if a sample cell is identified as an abnormal cell, the data of that sample cell will not be used in model training. In this embodiment, abnormal cells include cells with a missing sampling point ratio greater than a first threshold (hereinafter referred to as cells with more missing values), out-of-service cells, and broadcast testing cells. That is, if a cell belongs to at least one of the three types of cells, such as cells with more missing values, out-of-service cells, or broadcast testing cells, then the cell can be considered an abnormal cell. The first threshold can be set flexibly and is not limited in this application. For example, the first threshold can be 10%, 15%, etc.

[0125] In some embodiments, sampling data of the daily total traffic (uplink traffic + downlink traffic) metric can be used to identify abnormal cells. For example, to determine if a cell has a large number of missing values, the daily total traffic data for a cell over one year can be obtained. If the proportion of missing sampling points in this data to the total number of sampling points over one year (e.g., 365) is greater than a first threshold (e.g., 10%), then the cell is identified as having a large number of missing values ​​and is considered an abnormal cell. The same sampling data of the daily total traffic metric can be used to determine the status of out-of-service cells and cells undergoing broadcast testing.

[0126] For example, if a certain community has a monthly average traffic value of 0 for at least one month, then the community is judged to be a community that has been out of service. Figure 2 Example diagrams of daily total traffic index data for out-of-service cells and broadcast testing cells provided in this application embodiment, such as... Figure 2 As shown in the example on the left of the figure, the average monthly traffic of the cell has been 0 for more than 2 months, so the cell can be identified as a cell that has been out of service.

[0127] For example, if the ratio between the maximum monthly average traffic value and the minimum monthly average traffic value of a certain cell is greater than a second threshold, then that cell is a test cell. The second threshold can be set flexibly and is not limited in this application; for example, the second threshold could be 10, 20, etc. Figure 2 As shown in the figure, the maximum monthly average traffic value of the example cell on the right is much greater than the minimum monthly average traffic value of the cell, so the cell can be identified as a broadcast test cell.

[0128] If the cell to be identified is determined not to be an abnormal cell, then the subsequent processing of outlier sampling points (sampling points with abnormal indicator values) and missing sampling points (sampling points with only timestamps and empty indicator values) is performed. Failure to process outliers and missing values ​​in the data will affect the accuracy of the model's predictions; therefore, numerical processing is required for outlier and missing sampling points in the first processing data. Numerical processing refers to adjusting the indicator values ​​of outlier sampling points and filling in the indicator values ​​of missing sampling points, etc.

[0129] In some embodiments, numerical processing is performed on outlier and missing sample points in the first processed data, including:

[0130] Based on the Prophet model, the predicted data of various load characteristic indicators of the cell to be identified in the historical time period are obtained respectively.

[0131] Based on the predicted data and the preset confidence interval, identify the abnormal sampling points in the first processed data;

[0132] The predicted data is used to fill in the outlier and missing sample points in the first processed data.

[0133] Specifically, in this embodiment, the Prophet model is used to perform single-index prediction on each load characteristic index to obtain the predicted data of each load characteristic index of the cell to be identified in the historical time period. Then, these predicted data and preset confidence intervals are used to screen out abnormal sampling points.

[0134] The Prophet model analyzes various time series characteristics: periodicity, trend, holiday effects, and some outliers. Regarding trends, it supports incorporating points of change to achieve piecewise linear fitting. For periodicity, it uses Fourier series to build a periodic model (sin + cos). For holidays and sudden events, users can specify holidays and the N days preceding and following them via tables. The Prophet model can be considered an integrated solution for time series analysis. However, it can only predict single indicators, and its long-term predictions are not accurate enough. This application's embodiments utilize it for numerical processing of outlier and missing sampling points.

[0135] Taking downstream traffic as an example of a load characteristic indicator (other load characteristic indicators can be handled similarly for abnormal sampling points and missing sampling points), Figure 3 This is an example diagram showing the results of the Prophet model for downlink traffic prediction provided in this application embodiment. The curve in the diagram is the downlink traffic data curve of a certain cell within a certain period of time, fitted by the Prophet model. The solid circles in the diagram are the actual downlink traffic data scatter points within the same time period, the gray area is the confidence interval (for example, a confidence interval of 0.95 can be set), and the points outside the confidence interval are outliers, marked as hollow circles. Figure 3 The example prediction results use the following Prophet model parameter settings: holidays = statutory holidays, yearly_seasonality = 8, weekly_seasonality = 8, daily_seasonality = 20, changepoint_prior_scale = 0.04, growth = 'linear', interval_width = 0.95.

[0136] Then, for abnormal and missing sampling points, the predicted index values ​​with the same timestamp are used to fill in the numerical values ​​of the sampling points. For example, in the data of a certain load characteristic index of a certain community, the timestamp of an abnormal sampling point 1 is recorded as timestamp 1, and the timestamp of a missing sampling point 1 is recorded as timestamp 2. Then, the predicted index value of timestamp 1 in the predicted data of the load characteristic index of that community can be used to replace the original index value of abnormal sampling point 1, and the predicted index value of timestamp 2 in the predicted data of the load characteristic index of that community can be used as the index value of missing sampling point 1.

[0137] Figure 4 The following is a comparison diagram of downlink traffic abnormal sampling points and missing sampling points before and after processing, provided in the embodiments of this application. Figure 4 As shown in the figure, the processed data (represented by solid lines) is smoother and more complete than the unprocessed data (represented by hollow lines), which is beneficial to improving the accuracy of subsequent Informer model predictions.

[0138] Since the dimensions and orders of magnitude of various load characteristic indicators differ significantly, after numerically processing the abnormal and missing sampling points in the first processed data to obtain the second processed data of multiple load characteristic indicators, the second processed data can be standardized to constrain the data to a unified dimension in order to facilitate prediction by the Informer model.

[0139] Standardization refers to transforming raw data according to a certain ratio through specific mathematical transformations, so that it falls into a small, specific interval, such as [0, 1] or [-1, 1]. This eliminates the differences in characteristics, dimensions, and orders of magnitude between different variables, transforming them into a dimensionless relative value, also known as a standardized value. This ensures that the values ​​of all indicators are at the same order of magnitude, thus facilitating comprehensive analysis and comparison of indicators with different units or orders of magnitude.

[0140] In some embodiments, the standardization calculation formula is: Standardized value = (Original value - Mean) / Standard deviation. Here, the standardized value, original value, mean, and standard deviation are all specific to a single load characteristic indicator within a single cell. That is, during standardization, the mean and standard deviation are used for each load characteristic indicator. For example, for the standardization of uplink traffic data, the mean and standard deviation used are the mean and standard deviation of the cell's uplink traffic; for the standardization of uplink PRB average utilization rate data, the mean and standard deviation used are the mean and standard deviation of the uplink PRB average utilization rate.

[0141] Step 101: Input the time series data into the Informer model to obtain the multi-indicator prediction results output by the Informer model. The multi-indicator prediction results include the prediction results of multiple load characteristic indicators of the cell to be identified within the prediction time period.

[0142] The Informer model is trained based on multi-indicator historical time-series data from multiple sample cells. The multi-indicator historical time-series data includes historical time-series data of multiple load characteristic indicators.

[0143] Specifically, the Informer model is a time series prediction model improved based on the Transformer model. By employing probable sparse self-attention, self-attention distilling, and generative style decoder mechanisms, it reduces model complexity and memory usage, improves computational efficiency and inference speed for long-sequence prediction. Compared to the limitations of ARMA and Prophet models, which can only predict single indicators, and LSTM, which has limited effectiveness in predicting long periods, the Informer model performs well in predicting multiple indicators and both short and long periods.

[0144] Before identifying high-load cells, the Informer model is trained using historical time-series data of multiple load characteristics from a large number of sample cells. This historical time-series data includes historical time-series data of various load characteristics. The load characteristics used for Informer model training and inference are consistent. Historical time-series data for a specific load characteristic refers to the time-series data of that load characteristic over a past period (e.g., the past year).

[0145] By inputting the time-series data of multiple load characteristic indicators of the cell to be identified within a historical period into a pre-trained Informer model, the predicted results of multiple load characteristic indicators of the cell to be identified within the prediction period can be obtained. The prediction period is a certain future time period that the model needs to predict. The Informer model can perform long-term predictions (such as predicting the results of the next month) or short-term predictions (such as predicting the results of the next day).

[0146] Step 102: Based on the prediction results and the preset high load standard, determine whether the cell to be identified is a high load cell.

[0147] Specifically, after obtaining the prediction results, it is possible to determine whether the cell to be identified is a high-load cell based on the prediction results and the preset high-load criteria (the high-load criteria are set in accordance with load characteristic indicators). For example, if the prediction results of the cell to be identified meet the high-load criteria, then the cell to be identified can be determined as a high-load cell.

[0148] In some embodiments, the high load standard includes an uplink high load standard and a downlink high load standard;

[0149] The uplink high load criteria include: if the average uplink traffic of a cell is greater than the uplink traffic threshold and the average uplink PRB utilization rate of the cell is greater than the average uplink PRB utilization rate threshold during the forecast period, then the cell meets the uplink high load criteria.

[0150] The downlink high load criteria include: during the prediction period, the average downlink traffic of the cell is greater than the downlink traffic threshold, and at least one of the following is met: the average downlink PRB utilization rate of the cell is greater than the downlink PRB average utilization rate threshold, or the PDCCH channel CCE occupancy rate of the cell is greater than the PDCCH channel CCE occupancy rate threshold. In this case, the cell meets the downlink high load criteria.

[0151] Taking 5G cells as an example, the thresholds of various load characteristic indicators in the high load standard differ for cells with different coverage types, different operating frequency bands, different transceiver modes, and different main service package types. Table 1 below shows the high load standard thresholds for cells with different coverage types, different operating frequency bands, different transceiver modes, and different main service package types.

[0152] Table 1 Examples of high-load standard thresholds for different residential areas

[0153]

[0154]

[0155] The cell definitions for different packet types are shown in Table 2 below.

[0156] Table 2. Cell definitions for different packet types

[0157] Package type Flow per cell (KB) representing Quality of Service (QoS) Dabao Community >3000 Zhongbao Community (1500,3000] Xiaobao Community <＝1500

[0158] in,

[0159]

[0160] In some embodiments, determining whether a cell to be identified is a high-load cell based on prediction results and a preset high-load standard includes:

[0161] If the prediction results determine that the cell to be identified meets the uplink high load standard and / or downlink high load standard, then the cell to be identified is determined to be a high load cell.

[0162] Specifically, in this embodiment, when the high load standard includes both uplink high load standard and downlink high load standard, the cell is determined to be a high load cell as long as at least one of the uplink high load standard and downlink high load standard is met.

[0163] For example, assuming a cell is a macro base station, 700MHz, 4TR, large packet cell, according to the multi-indicator prediction results, its average uplink traffic for the next month is greater than the uplink traffic threshold of 2GB, and the average uplink PRB utilization rate is greater than the uplink PRB average utilization rate threshold of 50%, then the cell meets the uplink high load standard, and the cell can be determined as a high load cell.

[0164] For example, suppose a cell is a macro base station, 700MHz, 4TR, large packet cell. According to the multi-indicator prediction results, its average downlink traffic for the next month is greater than the downlink traffic threshold of 8GB, and it meets the following conditions: the average downlink PRB utilization rate is greater than the downlink PRB average utilization rate threshold of 70%, and / or the PDCCH channel CCE occupancy rate is greater than the PDCCH channel CCE occupancy rate threshold of 50%. Then the cell meets the downlink high load standard, and the cell can be determined as a high load cell.

[0165] For example, if a cell meets both the uplink high load standard and the downlink high load standard, it can also be determined that the cell is a high load cell.

[0166] The high-load cell intelligent identification method provided in this application uses the Informer model to predict multiple load characteristic indicators of the cell to be identified, and then determines the high-load cell based on the prediction results and the preset high-load standard. It can achieve intelligent identification of a large number of cells without the need for a large investment of human resources, which greatly improves the efficiency and accuracy of high-load cell identification.

[0167] In some embodiments, the encoder of the Informer model includes a multi-layer coding sub-module, each of which includes a multi-head probabilistic sparse self-attention sub-layer and a self-attention distillation sub-layer.

[0168] The Informer model's decoder consists of multiple decoding sub-modules, each of which includes a masked multi-head probabilistic sparse self-attention sub-layer and a multi-head attention sub-layer.

[0169] The encoder's output features are input into the decoder's mask multi-head probability sparse self-attention sublayer.

[0170] Specifically, the Informer model used in this embodiment is an improvement over existing Informer models. This Informer model comprises four parts: a feature processing module, an encoder, a decoder, and a fully connected layer. The core improvement lies in the encoder and decoder structures. The encoder of this Informer model includes multi-layer coding sub-modules, each including a multi-head probabilistic sparse self-attention sub-layer and a self-attention distillation sub-layer. The decoder includes multi-layer decoding sub-modules, each including a masked multi-head probabilistic sparse self-attention sub-layer and a multi-head attention sub-layer. Furthermore, unlike existing Informer models, the output features of the encoder in this Informer model are not the input multi-head self-attention sub-layer of the decoder, but rather the input masked multi-head probabilistic sparse self-attention sub-layer of the decoder. By using the improved Informer model, it is possible to adapt to multi-index prediction of massive cells, further improving the identification efficiency of high-load cells.

[0171] In some embodiments, the Informer model's encoder includes three encoding sub-modules, and the Informer model's decoder includes two decoding sub-modules. By appropriately setting the number of layers in the encoding and decoding sub-modules, the model complexity can be further reduced and the model inference speed improved while maintaining prediction accuracy.

[0172] In some embodiments, the training loss function of the Informer model includes the Log-Cosh loss function.

[0173] Specifically, the Log-Cosh loss function is a commonly used function in regression tasks. It is smoother than the L2 loss function, and the Log-Cosh loss function is the logarithm of the hyperbolic cosine of the prediction error. When the variable x is small, the loss function value log(cosh(x)) is approximately equal to (x... 2 When x is large, the loss function value log(cosh(x)) is approximately equal to abs(x) - log(2). This means that the "Log-Cosh" loss function works very similarly to MSE (mean squared error), but is not affected by occasional outliers. It has all the advantages of Huber loss and is differentiable in all cases. By training the Informer model using the Log-Cosh loss function, the trained Informer model can achieve better performance in multi-metric prediction.

[0174] Figure 5 The flowchart of Informer model training and prediction provided for embodiments of this application is as follows: Figure 5 As shown in this example, firstly, one year's worth of KPI data for a certain city is collected for model training. The collected data is first cleaned, then abnormal cells are removed, features are selected, and the Prophet algorithm is used to detect outliers and fill in missing values ​​for each load characteristic indicator. The data for each cell is then standardized, and finally, it is input into the improved Informer model. During prediction, the original data of multiple load characteristic indicators for the cell to be identified are first cleaned, then subjected to processes such as abnormal cell identification, outlier detection, missing value filling, and standardization. Finally, the obtained data is input into the improved Informer model to obtain the prediction results.

[0175] Figure 6 The Informer model structure and training example diagrams provided in the embodiments of this application are as follows: Figure 6 As shown in the figure, the structure of the improved Informer model is illustrated. Each encoding sub-module in the encoder of the improved Informer model consists of two sub-layers: a multi-head probabilistic sparse self-attention mechanism and a self-attention distillation mechanism. Each decoding sub-module in the decoder consists of two sub-layers: a mask multi-head probabilistic sparse self-attention mechanism and a multi-head attention mechanism. The improved Informer model uses the Log-Cosh loss function for backpropagation during training.

[0176] During prediction, preprocessed multi-indicator data of the cell to be identified is input into the improved Informer model to capture the long-term correlation between the input and output of the indicator sequences. Finally, the prediction result is obtained through a fully connected layer. The prediction results show that the improved Informer model provided in this application provides an error value of less than 0.3 for predicting indicators for the next month, with an average accuracy of 84% for each indicator, and an error value of less than 0.3 for predicting indicators for the next day, with an average accuracy of 89% for each indicator. The model performance is good.

[0177] The methods and apparatuses provided in the various embodiments of this application are based on the same concept. Since the methods and apparatuses solve problems in similar ways, the implementations of the apparatuses and methods can refer to each other, and repeated details will not be repeated.

[0178] Figure 7 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application, such as... Figure 7 As shown, the electronic device includes a memory 720, a transceiver 710, and a processor 700; wherein the processor 700 and the memory 720 can also be physically arranged separately.

[0179] The memory 720 is used to store computer programs; the transceiver 710 is used to send and receive data under the control of the processor 700.

[0180] Specifically, the transceiver 710 is used to receive and send data under the control of the processor 700.

[0181] Among them, Figure 7 In this application, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits of one or more processors represented by processor 700 and memory represented by memory 720 together. The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be further described herein. The bus interface provides an interface. The transceiver 710 can be multiple elements, including a transmitter and a receiver, providing a unit for communicating with various other devices over a transmission medium, including wireless channels, wired channels, optical fibers, and other transmission media.

[0182] The processor 700 is responsible for managing the bus architecture and general processing, while the memory 720 can store the data used by the processor 700 during operation.

[0183] The processor 700 can be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD). The processor can also adopt a multi-core architecture.

[0184] The processor 700 calls the computer program stored in the memory 720 to execute any of the methods provided in the embodiments of this application according to the obtained executable instructions, such as: acquiring time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period; inputting the time-series data into the Informer model to obtain the multi-indicator prediction results output by the Informer model, the multi-indicator prediction results including the prediction results of multiple load characteristic indicators of the cell to be identified within the prediction time period; determining whether the cell to be identified is a high-load cell based on the prediction results and the preset high-load standard; wherein, the Informer model is trained based on the multi-indicator historical time-series data of multiple sample cells, the multi-indicator historical time-series data including the historical time-series data of multiple load characteristic indicators.

[0185] In some embodiments, obtaining time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period includes:

[0186] The original data of multiple load characteristic indicators of the cell to be identified within a historical time period are cleaned to obtain the first processed data of multiple load characteristic indicators.

[0187] Based on the first processed data, it is determined whether the cell to be identified is an abnormal cell; abnormal cells include cells with a missing sampling point ratio greater than the first threshold, out-of-service cells, and broadcast test cells;

[0188] If it is determined that the cell to be identified is not an abnormal cell, the abnormal sampling points and missing sampling points in the first processed data are numerically processed to obtain the second processed data with multiple load characteristic indicators.

[0189] The second processing data is standardized to obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical period.

[0190] In some embodiments, numerical processing is performed on outlier and missing sample points in the first processed data, including:

[0191] Based on the Prophet model, the predicted data of various load characteristic indicators of the cell to be identified in the historical time period are obtained respectively.

[0192] Based on the predicted data and the preset confidence interval, identify the abnormal sampling points in the first processed data;

[0193] The predicted data is used to fill in the outlier and missing sample points in the first processed data.

[0194] In some embodiments, the method for determining a cell that has gone out of service includes:

[0195] If a community has a monthly average traffic value of 0 for at least one month, then the community is considered to be out of service.

[0196] In some embodiments, the method for determining the broadcast test cell includes:

[0197] If the ratio between the maximum monthly average traffic value and the minimum monthly average traffic value of a cell is greater than the second threshold, then the cell is a broadcast test cell.

[0198] In some embodiments, load characteristics include uplink traffic of the cell, downlink traffic of the cell, average uplink physical resource block (PRB) utilization of the cell, average downlink PRB utilization of the cell, and physical downlink control channel (PDCCH) control channel element (CCE) occupancy rate of the cell.

[0199] In some embodiments, the high load standard includes an uplink high load standard and a downlink high load standard;

[0200] The uplink high load criteria include: if the average uplink traffic of the cell is greater than the uplink traffic threshold and the average uplink PRB utilization rate of the cell is greater than the average uplink PRB utilization rate threshold during the forecast period, then the cell meets the uplink high load criteria.

[0201] The downlink high load criteria include: during the prediction period, the average downlink traffic of the cell is greater than the downlink traffic threshold, and at least one of the following is met: the average downlink PRB utilization rate of the cell is greater than the downlink PRB average utilization rate threshold, or the PDCCH channel CCE occupancy rate of the cell is greater than the PDCCH channel CCE occupancy rate threshold.

[0202] In some embodiments, determining whether a cell to be identified is a high-load cell based on prediction results and a preset high-load standard includes:

[0203] If the prediction results determine that the cell to be identified meets the uplink high load standard and / or downlink high load standard, then the cell to be identified is determined to be a high load cell.

[0204] In some embodiments, the encoder of the Informer model includes a multi-layer coding sub-module, each of which includes a multi-head probabilistic sparse self-attention sub-layer and a self-attention distillation sub-layer.

[0205] The Informer model's decoder consists of multiple decoding sub-modules, each of which includes a masked multi-head probabilistic sparse self-attention sub-layer and a multi-head attention sub-layer.

[0206] The encoder's output features are input into the decoder's mask multi-head probability sparse self-attention sublayer.

[0207] In some embodiments, the training loss function of the Informer model includes the Log-Cosh loss function.

[0208] In some embodiments, the encoder of the Informer model includes a 3-layer encoding submodule, and the decoder of the Informer model includes a 2-layer decoding submodule.

[0209] It should be noted that the electronic device provided in this application embodiment can implement all the method steps implemented in the above method embodiment and can achieve the same technical effect. Here, the parts and beneficial effects that are the same as those in the method embodiment will not be described in detail.

[0210] Figure 8 This is a schematic diagram of the structure of the high-load cell intelligent identification device provided in the embodiments of this application, as shown below. Figure 8 As shown, the device includes:

[0211] The acquisition unit 800 is used to acquire time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period;

[0212] The prediction unit 810 is used to input time series data into the Informer model and obtain the multi-indicator prediction results output by the Informer model. The multi-indicator prediction results include the prediction results of multiple load characteristic indicators of the cell to be identified within the prediction time period.

[0213] The determination unit 820 is used to determine whether the cell to be identified is a high-load cell based on the prediction results and the preset high-load criteria.

[0214] The Informer model is trained based on multi-indicator historical time-series data from multiple sample cells. The multi-indicator historical time-series data includes historical time-series data of multiple load characteristic indicators.

[0215] In some embodiments, obtaining time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period includes:

[0216] The original data of multiple load characteristic indicators of the cell to be identified within a historical time period are cleaned to obtain the first processed data of multiple load characteristic indicators.

[0217] Based on the first processed data, it is determined whether the cell to be identified is an abnormal cell; abnormal cells include cells with a missing sampling point ratio greater than the first threshold, out-of-service cells, and broadcast test cells;

[0218] If it is determined that the cell to be identified is not an abnormal cell, the abnormal sampling points and missing sampling points in the first processed data are numerically processed to obtain the second processed data with multiple load characteristic indicators.

[0219] The second processing data is standardized to obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical period.

[0220] In some embodiments, numerical processing is performed on outlier and missing sample points in the first processed data, including:

[0221] Based on the Prophet model, the predicted data of various load characteristic indicators of the cell to be identified in the historical time period are obtained respectively.

[0222] Based on the predicted data and the preset confidence interval, identify the abnormal sampling points in the first processed data;

[0223] The predicted data is used to fill in the outlier and missing sample points in the first processed data.

[0224] In some embodiments, the method for determining a cell that has gone out of service includes:

[0225] If a community has a monthly average traffic value of 0 for at least one month, then the community is considered to be out of service.

[0226] In some embodiments, the method for determining the broadcast test cell includes:

[0227] If the ratio between the maximum monthly average traffic value and the minimum monthly average traffic value of a cell is greater than the second threshold, then the cell is a broadcast test cell.

[0228] In some embodiments, load characteristics include uplink traffic of the cell, downlink traffic of the cell, average uplink physical resource block (PRB) utilization of the cell, average downlink PRB utilization of the cell, and physical downlink control channel (PDCCH) control channel element (CCE) occupancy rate of the cell.

[0229] In some embodiments, the high load standard includes an uplink high load standard and a downlink high load standard;

[0230] The uplink high load criteria include: if the average uplink traffic of the cell is greater than the uplink traffic threshold and the average uplink PRB utilization rate of the cell is greater than the average uplink PRB utilization rate threshold during the forecast period, then the cell meets the uplink high load criteria.

[0231] The downlink high load criteria include: during the prediction period, the average downlink traffic of the cell is greater than the downlink traffic threshold, and at least one of the following is met: the average downlink PRB utilization rate of the cell is greater than the downlink PRB average utilization rate threshold, or the cell's PDCCH channel CCE occupancy rate is greater than the PDCCH channel CCE occupancy rate threshold.

[0232] In some embodiments, determining whether a cell to be identified is a high-load cell based on prediction results and a preset high-load standard includes:

[0233] If the prediction results determine that the cell to be identified meets the uplink high load standard and / or downlink high load standard, then the cell to be identified is determined to be a high load cell.

[0234] In some embodiments, the encoder of the Informer model includes a multi-layer coding sub-module, each of which includes a multi-head probabilistic sparse self-attention sub-layer and a self-attention distillation sub-layer.

[0235] The Informer model's decoder consists of multiple decoding sub-modules, each of which includes a masked multi-head probabilistic sparse self-attention sub-layer and a multi-head attention sub-layer.

[0236] The encoder's output features are input into the decoder's mask multi-head probability sparse self-attention sublayer.

[0237] In some embodiments, the training loss function of the Informer model includes the Log-Cosh loss function.

[0238] In some embodiments, the encoder of the Informer model includes a 3-layer encoding submodule, and the decoder of the Informer model includes a 2-layer decoding submodule.

[0239] It should be noted that the division of units in the embodiments of this application is illustrative and only represents one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units.

[0240] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a processor-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0241] It should be noted that the apparatus provided in this application embodiment can implement all the method steps implemented in the above method embodiment and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiment and the beneficial effects will not be described in detail.

[0242] On the other hand, embodiments of this application also provide a non-transiently readable storage medium storing a computer program, which is used to cause a processor to execute the high-load cell intelligent identification method provided in the above embodiments.

[0243] It should be noted that the non-transiently readable storage medium provided in this application embodiment can implement all the method steps implemented in the above method embodiment and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiment and the beneficial effects will not be described in detail.

[0244] The non-transiently readable storage medium can be any available medium or data storage device that a computer can access, including but not limited to magnetic storage (e.g., floppy disks, hard disks, magnetic tapes, magneto-optical disks (MOs), etc.), optical storage (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0245] The technical solutions provided in this application can be applied to various systems, especially 5G systems. For example, applicable systems may include Global System for Mobile Communication (GSM), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA) General Packet Radio Service (GPRS), Long Term Evolution (LTE), LTE Frequency Division Duplex (FDD), LTE Time Division Duplex (TDD), Long Term Evolution Advanced (LTE-A), Universal Mobile Telecommunication System (UMTS), Worldwide Interoperability for Microwave Access (WiMAX), and 5G New Radio (NR). All of these systems include terminal equipment and network equipment. The systems may also include a core network component, such as Evolved Packet System (EPS) and 5G system (5GS).

[0246] The network device involved in this application embodiment can be a base station, which may include multiple cells providing services to terminals. Depending on the specific application, a base station may also be called an access point, or a device in an access network that communicates with a wireless terminal device through one or more sectors on the air interface, or other names. The network device can be used to exchange received air frames with Internet Protocol (IP) packets, acting as a router between the wireless terminal device and the rest of the access network, where the rest of the access network may include an Internet Protocol (IP) communication network. The network device can also coordinate the attribute management of the air interface. For example, the network equipment involved in the embodiments of this application can be a base transceiver station (BTS) in a Global System for Mobile communications (GSM) or Code Division Multiple Access (CDMA), a NodeB in a Wide-band Code Division Multiple Access (WCDMA) system, an evolved Node B (eNB or e-NodeB) in a long term evolution (LTE) system, a 5G base station (gNB) in a next generation system, a Home evolved Node B (HeNB), a relay node, a femto, a pico, etc., and is not limited in the embodiments of this application. In some network structures, the network equipment may include centralized unit (CU) nodes and distributed unit (DU) nodes, and the centralized unit and distributed unit may be geographically separated.

[0247] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0248] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-executable instructions. These computer-executable instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0249] These processor-executable instructions may also be stored in a processor-readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the processor-readable memory produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0250] These processors can execute instructions that can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0251] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for intelligent identification of high-load cells, characterized in that, include: Obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical period; The time series data is input into the Informer model to obtain the multi-indicator prediction results output by the Informer model. The multi-indicator prediction results include the prediction results of the multiple load characteristic indicators of the cell to be identified within the prediction time period. Based on the prediction results and the preset high-load criteria, determine whether the cell to be identified is a high-load cell; The Informer model is trained based on historical time-series data of multiple sample cells, including historical time-series data of the multiple load characteristic indicators.

2. The intelligent identification method for high-load cells according to claim 1, characterized in that, The acquisition of time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period includes: The original data of multiple load characteristic indicators of the cell to be identified within a historical time period are cleaned to obtain the first processed data of the multiple load characteristic indicators. Based on the first processed data, it is determined whether the cell to be identified is an abnormal cell; the abnormal cell includes cells with a missing sampling point ratio greater than a first threshold, out-of-service cells, and broadcast test cells; If it is determined that the cell to be identified is not an abnormal cell, the abnormal sampling points and missing sampling points in the first processed data are numerically processed to obtain the second processed data of the multiple load characteristic indicators. The second processed data is standardized to obtain time-series data of the multiple load characteristic indicators of the cell to be identified within the historical time period.

3. The intelligent identification method for high-load cells according to claim 2, characterized in that, The numerical processing of abnormal and missing sampling points in the first processed data includes: Based on the Prophet model, the predicted data of each load characteristic index of the cell to be identified within the historical time period are obtained respectively. Based on the predicted data and the preset confidence interval, determine the abnormal sampling points in the first processed data; The predicted data is used to fill in the abnormal and missing sampling points in the first processed data.

4. The intelligent identification method for high-load cells according to claim 2, characterized in that, The methods for determining the out-of-service cells include: If a cell has a monthly average traffic value of 0 for at least one month, then the cell is considered to be out of service.

5. The intelligent identification method for high-load cells according to claim 2, characterized in that, The methods for determining the broadcast test cells include: If the ratio between the maximum monthly average traffic value and the minimum monthly average traffic value of a cell is greater than a second threshold, then the cell is a broadcast test cell.

6. The high-load cell intelligent identification method according to any one of claims 1 to 3, characterized in that, The load characteristic indicators include the cell's uplink traffic, the cell's downlink traffic, the cell's average uplink physical resource block (PRB) utilization rate, the cell's average downlink PRB utilization rate, and the cell's physical downlink control channel (PDCCH) control channel element (CCE) occupancy rate.

7. The intelligent identification method for high-load cells according to claim 6, characterized in that, The high load standard includes an uplink high load standard and a downlink high load standard; The uplink high load standard includes: if the average uplink traffic of the cell is greater than the uplink traffic threshold during the predicted time period, and the average uplink PRB utilization rate of the cell is greater than the average uplink PRB utilization rate threshold, then the cell meets the uplink high load standard. The downlink high load criterion includes: during the predicted time period, the average downlink traffic of the cell is greater than the downlink traffic threshold, and at least one of the following is met: the average downlink PRB utilization rate of the cell is greater than the downlink PRB average utilization rate threshold, and the PDCCH channel CCE occupancy rate of the cell is greater than the PDCCH channel CCE occupancy rate threshold, then the cell meets the downlink high load criterion.

8. The intelligent identification method for high-load cells according to claim 7, characterized in that, The step of determining whether the cell to be identified is a high-load cell based on the prediction result and the preset high-load standard includes: If the cell to be identified is determined to meet the uplink high load standard and / or the downlink high load standard based on the prediction results, then the cell to be identified is determined to be a high load cell.

9. The intelligent identification method for high-load cells according to claim 1, characterized in that, The encoder of the Informer model includes multiple coding sub-modules, each of which includes a multi-head probabilistic sparse self-attention sub-layer and a self-attention distillation sub-layer. The decoder of the Informer model includes multiple decoding sub-modules, each of which includes a masked multi-head probabilistic sparse self-attention sub-layer and a multi-head attention sub-layer. The output features of the encoder are input into the masked multi-head probabilistic sparse self-attention sublayer of the decoder.

10. The intelligent identification method for high-load cells according to claim 9, characterized in that, The training loss function of the Informer model includes the Log-Cosh loss function.

11. The intelligent identification method for high-load cells according to claim 9, characterized in that, The encoder of the Informer model includes three layers of encoding sub-modules, and the decoder of the Informer model includes two layers of decoding sub-modules.

12. An electronic device, characterized in that, Includes memory, transceiver, and processor; A memory for storing computer programs; a transceiver for sending and receiving data under the control of the processor; and a processor for reading the computer programs from the memory and performing the following operations: Obtain time-series data of multiple load characteristic indicators of the cell to be identified within a historical period; The time series data is input into the Informer model to obtain the multi-indicator prediction results output by the Informer model. The multi-indicator prediction results include the prediction results of the multiple load characteristic indicators of the cell to be identified within the prediction time period. Based on the prediction results and the preset high-load criteria, determine whether the cell to be identified is a high-load cell; The Informer model is trained based on historical time-series data of multiple sample cells, including historical time-series data of the multiple load characteristic indicators.

13. The electronic device according to claim 12, characterized in that, The acquisition of time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period includes: The original data of multiple load characteristic indicators of the cell to be identified within a historical time period are cleaned to obtain the first processed data of the multiple load characteristic indicators. Based on the first processed data, it is determined whether the cell to be identified is an abnormal cell; the abnormal cell includes cells with a missing sampling point ratio greater than a first threshold, out-of-service cells, and broadcast test cells; If it is determined that the cell to be identified is not an abnormal cell, the abnormal sampling points and missing sampling points in the first processed data are numerically processed to obtain the second processed data of the multiple load characteristic indicators. The second processed data is standardized to obtain time-series data of the multiple load characteristic indicators of the cell to be identified within the historical time period.

14. The electronic device according to claim 13, characterized in that, The numerical processing of abnormal and missing sampling points in the first processed data includes: Based on the Prophet model, the predicted data of each load characteristic index of the cell to be identified within the historical time period are obtained respectively. Based on the predicted data and the preset confidence interval, determine the abnormal sampling points in the first processed data; The predicted data is used to fill in the abnormal and missing sampling points in the first processed data.

15. The electronic device according to claim 13, characterized in that, The methods for determining the out-of-service cells include: If a cell has a monthly average traffic value of 0 for at least one month, then the cell is considered to be out of service.

16. The electronic device according to claim 13, characterized in that, The methods for determining the broadcast test cells include: If the ratio between the maximum monthly average traffic value and the minimum monthly average traffic value of a cell is greater than a second threshold, then the cell is a broadcast test cell.

17. The electronic device according to any one of claims 12 to 14, characterized in that, The load characteristic indicators include the cell's uplink traffic, the cell's downlink traffic, the cell's average uplink physical resource block (PRB) utilization rate, the cell's average downlink PRB utilization rate, and the cell's physical downlink control channel (PDCCH) control channel element (CCE) occupancy rate.

18. The electronic device according to claim 17, characterized in that, The high load standard includes an uplink high load standard and a downlink high load standard; The uplink high load standard includes: if the average uplink traffic of the cell is greater than the uplink traffic threshold during the predicted time period, and the average uplink PRB utilization rate of the cell is greater than the average uplink PRB utilization rate threshold, then the cell meets the uplink high load standard. The downlink high load criterion includes: during the predicted time period, the average downlink traffic of the cell is greater than the downlink traffic threshold, and at least one of the following is met: the average downlink PRB utilization rate of the cell is greater than the downlink PRB average utilization rate threshold, and the PDCCH channel CCE occupancy rate of the cell is greater than the PDCCH channel CCE occupancy rate threshold, then the cell meets the downlink high load criterion.

19. The electronic device according to claim 18, characterized in that, The step of determining whether the cell to be identified is a high-load cell based on the prediction result and the preset high-load standard includes: If the cell to be identified is determined to meet the uplink high load standard and / or the downlink high load standard based on the prediction results, then the cell to be identified is determined to be a high load cell.

20. The electronic device according to claim 12, characterized in that, The encoder of the Informer model includes multiple coding sub-modules, each of which includes a multi-head probabilistic sparse self-attention sub-layer and a self-attention distillation sub-layer. The decoder of the Informer model includes multiple decoding sub-modules, each of which includes a masked multi-head probabilistic sparse self-attention sub-layer and a multi-head attention sub-layer. The output features of the encoder are input into the masked multi-head probabilistic sparse self-attention sublayer of the decoder.

21. The electronic device according to claim 20, characterized in that, The training loss function of the Informer model includes the Log-Cosh loss function.

22. The electronic device according to claim 20, characterized in that, The encoder of the Informer model includes three layers of encoding sub-modules, and the decoder of the Informer model includes two layers of decoding sub-modules.

23. A high-load cell intelligent identification device, characterized in that, include: The acquisition unit is used to acquire time-series data of multiple load characteristic indicators of the cell to be identified within a historical time period; The prediction unit is used to input the time series data into the Informer model and obtain the multi-index prediction results output by the Informer model. The multi-index prediction results include the prediction results of the multiple load characteristic indicators of the cell to be identified within the prediction time period. The determining unit is used to determine whether the cell to be identified is a high-load cell based on the prediction results and the preset high-load standard. The Informer model is trained based on historical time-series data of multiple sample cells, including historical time-series data of the multiple load characteristic indicators.

24. A non-transiently readable storage medium, characterized in that, The non-transiently readable storage medium stores a computer program that causes a processor to perform the method according to any one of claims 1 to 11.

Images