Cell maintenance priority determination method and device, electronic equipment and storage medium

Abstract

The invention provides a cell maintenance priority determination method and device, electronic equipment and a storage medium, and relates to the technical field of communication, and the method comprises the steps: obtaining a plurality of pieces of complaint information and cell unique identifiers ECI of a plurality of degraded cells; the ECI of at least one first cell associated with the multiple pieces of complaint information is determined, the at least one first cell is set as a first maintenance priority, and the at least one first cell is a cell in the multiple degraded cells; evaluation information of each second cell in a plurality of second cells is acquired, the evaluation information comprises a degradation index and / or historical data corresponding to the degradation index, and the plurality of second cells are cells except the at least one first cell in the plurality of degraded cells; and calculating a second maintenance priority of each second cell based on the evaluation information, wherein the second maintenance priority is lower than the first maintenance priority. According to the invention, the use experience of the user is improved by determining the maintenance priority.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and in particular, to a method, apparatus, electronic device, and storage medium for determining the maintenance priority of a cell. Background Art

[0002] The maintenance of a communication cell is an important means to maintain the normal operation of the equipment in the communication cell. In the related art, cell maintenance is usually regular maintenance or temporary maintenance. Among them, when there are user complaints about the cell or abnormal device parameters, the operation and maintenance personnel perform temporary maintenance on the cell. However, in the related art, when there are a large number of cells that need temporary maintenance, the maintenance personnel cannot confirm the parameter degradation situation of different cells, and there is a situation where a cell with slightly degraded parameters is maintained first, and a cell with severely degraded parameters is maintained later. The cell with severely degraded parameters is not processed in time, resulting in a poor user experience of the cell.

[0003] It can be seen that in the related art, there is a problem that the cell with severely degraded parameters is not processed in time, resulting in a poor user experience of the cell. Summary of the Invention

[0004] Embodiments of the present invention provide a method, apparatus, electronic device, and storage medium for determining the maintenance priority of a cell to solve the problem in the related art that the cell with severely degraded parameters is not processed in time, resulting in a poor user experience of the cell.

[0005] To solve the above problems, the present invention is implemented as follows:

[0006] In a first aspect, an embodiment of the present invention provides a method for determining the maintenance priority of a cell, including:

[0007] Obtain a plurality of complaint information and the cell unique identifier ECI of a plurality of degraded cells;

[0008] Determine the ECI of at least one first cell associated with the plurality of complaint information, and set the at least one first cell as the first maintenance priority, where the at least one first cell is a cell among the plurality of degraded cells;

[0009] Obtain the evaluation information of each second cell in a plurality of second cells, where the evaluation information includes a degradation index and / or historical data corresponding to the degradation index, and the plurality of second cells are cells among the plurality of degraded cells except the at least one first cell;

[0010] Calculate the second maintenance priority of each second cell based on the evaluation information, where the second maintenance priority is lower than the first maintenance priority.

[0011] In a second aspect, an embodiment of the present invention further provides a device for determining the maintenance priority of a cell, including:

[0012] A first acquisition module, configured to acquire a plurality of complaint information and the cell unique identifier ECI of a plurality of degraded cells;

[0013] A determination module, configured to determine the ECI of at least one first cell associated with the plurality of complaint information, and set the at least one first cell to a first maintenance priority, where the at least one first cell is a cell among the plurality of degraded cells;

[0014] A second acquisition module, configured to acquire the evaluation information of each second cell in a plurality of second cells, where the evaluation information includes a degradation index and / or historical data corresponding to the degradation index, and the plurality of second cells are cells among the plurality of degraded cells other than the at least one first cell;

[0015] A calculation module, configured to calculate the second maintenance priority of each second cell based on the evaluation information, where the second maintenance priority is lower than the first maintenance priority.

[0016] In a third aspect, an embodiment of the present invention further provides an electronic device, including a transceiver and a processor,

[0017] The transceiver is configured to acquire a plurality of complaint information and the cell unique identifier ECI of a plurality of degraded cells;

[0018] The processor is configured to determine the ECI of at least one first cell associated with the plurality of complaint information, and set the at least one first cell to a first maintenance priority, where the at least one first cell is a cell among the plurality of degraded cells;

[0019] The transceiver is further configured to acquire the evaluation information of each second cell in a plurality of second cells, where the evaluation information includes a degradation index and / or historical data corresponding to the degradation index, and the plurality of second cells are cells among the plurality of degraded cells other than the at least one first cell;

[0020] The processor is further configured to calculate the second maintenance priority of each second cell based on the evaluation information, where the second maintenance priority is lower than the first maintenance priority.

[0021] In a fourth aspect, an embodiment of the present invention provides an electronic device, including: a processor, a memory, and a program stored on the memory and executable on the processor, where when the program is executed by the processor, the steps of the cell maintenance priority determination method described in the first aspect are implemented.

[0022] Fifth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the cell maintenance priority determination method described in the first aspect above are implemented.

[0023] Sixth aspect, the present invention further provides a computer program product, including computer instructions. When the computer instructions are executed by a processor, the steps in the cell maintenance priority determination method described in the first aspect above are implemented.

[0024] In an embodiment of the present invention, a plurality of complaint information and the cell unique identifier ECI of a plurality of degraded cells are obtained; the ECI of at least one first cell associated with the plurality of complaint information is determined, and the at least one first cell is set as the first maintenance priority, and the at least one first cell is a cell among the plurality of degraded cells; the evaluation information of each second cell in a plurality of second cells is obtained, where the evaluation information includes a degradation index and / or historical data corresponding to the degradation index, and the plurality of second cells are cells among the plurality of degraded cells other than the at least one first cell; based on the evaluation information, the second maintenance priority of each second cell is calculated, and the second maintenance priority is lower than the first maintenance priority. In this way, the first maintenance priority or the second maintenance priority corresponding to a plurality of degraded cells is determined through the complaint information and the evaluation information, so that the maintenance personnel can maintain the plurality of degraded cells in sequence according to the first priority and the second priority, avoiding the situation that the cells with serious parameter degradation are not processed in time, thereby effectively improving the usage experience of the cells. Description of the Drawings

[0025] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments of the present invention. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0026] Figure 1 is a flowchart of a cell maintenance priority determination method provided by an embodiment of the present invention;

[0027] Figure 2 is a schematic flowchart of maintenance priority determination provided by an embodiment of the present invention;

[0028] Figure 3 is a schematic flowchart of the association between complaint information and the first cell provided by an embodiment of the present invention;

[0029] Figure 4 is a schematic flowchart of calculating the complaint probability provided by an embodiment of the present invention;

[0030] Figure 5 It is a structural diagram of a cell maintenance priority determination device provided by an embodiment of the present invention;

[0031] Figure 6 It is a structural diagram of an electronic device provided by an embodiment of the present invention. Specific embodiments

[0032] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0033] The embodiments of the present invention provide a method, device, electronic device and storage medium for determining the cell maintenance priority. By determining the maintenance priorities of different deteriorated cells, the cells with severe deterioration are preferentially maintained, solving the problem in the related art that the cells with severely deteriorated parameters are not processed in time, resulting in poor usage experience of the cells.

[0034] Please refer to Figure 1 , Figure 1 It is a flowchart of a method for determining the cell maintenance priority provided by an embodiment of the present invention. As Figure 1 shown, it includes the following steps:

[0035] Step 101, obtain a plurality of complaint information and the cell unique identifier (Enhanced Cell Identity, ECI) of a plurality of deteriorated cells.

[0036] The above-mentioned plurality of complaint information is the information for users to complain about the services of the cell, such as complaining about call quality, complaining about network signal, etc. It should be noted that the plurality of complaint information is the complaint information within a period of time, and this period of time is the period after the cell indicators deteriorate. If there is complaint information before the cell indicators deteriorate, it is considered that this complaint information has no association with the cell indicator deterioration.

[0037] For each of the above-mentioned plurality of deteriorated cells, at least one indicator has deteriorated. Among them, the indicator is an indicator in a preset parameter set, and the preset parameter set is a pre-configured parameter set, including a plurality of parameter indicators that need to be monitored, such as wireless network connection rate, wireless network disconnection rate, cell handover success rate, etc.

[0038] Step 102: Determine the ECI of at least one first cell associated with the multiple complaint messages, and set the at least one first cell to the first maintenance priority level. The at least one first cell is a cell among the multiple deteriorated cells.

[0039] The ECI of the at least one first cell is associated with the multiple complaint messages, that is, at least one complaint message corresponds to one first cell. In this case, the deterioration of the indicators of the first cell has affected the user experience and needs to be processed first. At this time, the at least one first cell is set to the first maintenance priority level, so that the maintenance personnel can give priority to processing the cells with the first maintenance priority level, and the first cells with serious indicator deterioration can be maintained in time.

[0040] Step 103: Obtain the evaluation information of each second cell among the multiple second cells. The evaluation information includes deteriorated indicators and / or historical data corresponding to the deteriorated indicators. The multiple second cells are the cells among the multiple deteriorated cells other than the at least one first cell.

[0041] The multiple second cells are the cells that are not associated with complaint messages, and their maintenance priority levels are lower than those of the first cells. However, since there are still deteriorated indicators in the second cells, to prevent the long-term deterioration of the indicators in the second cells from affecting the user experience, the maintenance personnel still need to maintain the second cells.

[0042] The above-mentioned evaluation information includes deteriorated indicators and / or historical data corresponding to the deteriorated indicators. Among them, the deteriorated indicators can directly reflect the deterioration of the indicators of the first cell, and the historical data corresponding to the deteriorated indicators can reflect the possibility of further deterioration of the indicators. By obtaining the deteriorated indicators and / or the historical data corresponding to the deteriorated indicators, the deterioration of the indicators can be evaluated from one or more aspects of the deteriorated indicators and historical data, so as to determine the maintenance priority level of the second cell.

[0043] Step 104: Calculate the second maintenance priority level of each second cell based on the evaluation information. The second maintenance priority level is lower than the first maintenance priority level.

[0044] It should be noted that the evaluation information includes deteriorated indicators and / or historical data corresponding to the deteriorated indicators. Calculating the second maintenance priority level of each second cell based on the evaluation information can be calculating the second maintenance priority level of each second cell based on the deteriorated indicators, or calculating the second maintenance priority level of each second cell based on the historical data corresponding to the deteriorated indicators, or calculating the second maintenance priority level of each second cell based on the deteriorated indicators and the historical data corresponding to the deteriorated indicators, so as to evaluate the deterioration of the indicators from multiple aspects and obtain a more accurate second maintenance priority level.

[0045] Among them, since the second maintenance priority is the maintenance priority of the second cell, and the second cell is not associated with any complaint information, the second maintenance priority of the second cell is lower than the first maintenance priority of the first cell, so as to ensure that the cell with complaints is maintained first.

[0046] In an embodiment of the present invention, a plurality of complaint information and identifiers of a plurality of deteriorated cells are obtained; at least one first cell associated with the plurality of complaint information is determined, and the at least one first cell is set as the first maintenance priority, and the at least one first cell is a cell among the plurality of deteriorated cells; evaluation information of each second cell in the plurality of second cells is obtained, the evaluation information includes a deterioration index and / or historical data corresponding to the deterioration index, and the plurality of second cells are cells other than the at least one first cell among the plurality of deteriorated cells; the second maintenance priority of each second cell is calculated based on the evaluation information, and the second maintenance priority is lower than the first maintenance priority. In this way, the first maintenance priority or the second maintenance priority corresponding to a plurality of deteriorated cells is determined through the complaint information and the evaluation information, so that maintenance personnel can maintain the plurality of deteriorated cells in sequence according to the first priority and the second priority, avoiding the situation that cells with serious parameter deterioration are not processed in time, thereby effectively improving the usage experience of the cells.

[0047] In some embodiments, after calculating the second maintenance priority of each second cell based on the evaluation information, the method further includes:

[0048] Sending the first maintenance priority or the second maintenance priority corresponding to the plurality of cells to the terminal of the maintenance personnel, so that the maintenance personnel can maintain the plurality of deteriorated cells in sequence based on the first maintenance priority and the second maintenance priority.

[0049] Further, the process of maintaining the deteriorated cell is as Figure 2 shown. After obtaining the input data (a plurality of complaint information and identifiers of a plurality of deteriorated cells), the complaint information is associated with the deteriorated cells; in the case where a deteriorated cell is associated with complaint information, the deteriorated cell is set as the first maintenance priority; in the case where a deteriorated cell is not associated with complaint information, the complaint probability of the deteriorated cell being complained is determined through the historical data of the deterioration index of the deteriorated cell, and the deterioration score is determined through the deterioration index of the deteriorated cell, and then the second maintenance priority of the deteriorated cell is determined based on the complaint probability and the deterioration score.

[0050] In one embodiment, each piece of complaint information in the plurality of complaint information includes first location information, and determining the ECI of at least one first cell associated with the plurality of complaint information includes:

[0051] Obtain the second location information corresponding to multiple base stations, where each base station among the multiple base stations includes at least one degraded cell, and different degraded cells correspond to different azimuth ranges of the base station;

[0052] Based on the first location information and the second location information, calculate the distances between the locations where each complaint information is located and the multiple base stations;

[0053] Set the base station with the minimum distance corresponding to the target complaint information as the associated base station of the target complaint information, where the target complaint information is one of the multiple complaint information;

[0054] Calculate the first azimuth between the location where the target complaint information is located and the associated base station;

[0055] Based on the first azimuth, determine the ECI of the first cell associated with the target complaint information, where the first cell is a cell included in the associated base station, and the azimuth range corresponding to the first cell includes the first azimuth.

[0056] It should be noted that when a user makes a complaint, the complaint information sent includes the first location information of the user's location, and the specific degraded cell associated with the complaint information is determined through the first location information.

[0057] In an embodiment of the present invention, obtain the second location information corresponding to multiple base stations, where each base station among the multiple base stations includes at least one degraded cell, and different degraded cells correspond to different azimuth ranges of the base station; based on the first location information and the second location information, calculate the distances between the locations where each complaint information is located and the multiple base stations; set the base station with the minimum distance corresponding to the target complaint information as the associated base station of the target complaint information, where the target complaint information is one of the multiple complaint information; calculate the first azimuth between the location where the target complaint information is located and the associated base station; based on the first azimuth, determine the first cell associated with the target complaint information, where the first cell is a cell included in the associated base station, and the azimuth range corresponding to the first cell includes the first azimuth. In this way, through the first location information and the second location information, the complaint information is first associated with the base station, and then the cell associated with the complaint information is determined through the first azimuth, so as to realize the association between the complaint information and the cell.

[0058] In some embodiments, the first location information is the longitude and latitude information of the location where the complaint information is located, which can be expressed as (x ai , y ai ); the second location information is the longitude and latitude information of the base station, which can be expressed as [(x1, y1), (x2, y2), …, (x m , y m)], where m is the number of base stations; based on the first location information and the second location information, calculate the minimum distance between the location where the target complaint information is located and the base station (x d , y d ) can be expressed by the following formula:

[0059]

[0060] Calculate the minimum distance corresponding to the target complaint information through the above formula, and then the associated base station corresponding to the target complaint information can be determined.

[0061] In some embodiments, the first direction angle is the included angle between the location where the complaint information is located and the true north relative to the base station. Calculate the first direction angle between the location where the target complaint information is located and the associated base station, which can be calculated by the following formula:

[0062]

[0063] Calculate the value of the first direction angle through the above formula.

[0064] Furthermore, since the value range of the arctan function is Therefore, it is necessary to perform angle conversion on the calculated included angle to obtain the correct first direction angle, and the angle conversion process can be expressed by the following formula:

[0065]

[0066] Perform angle conversion through the above formula to obtain the accurate first direction angle.

[0067] In some embodiments, one base station corresponds to three cells, and the direction angle corresponding to each cell is The corresponding direction angle range is ±60° of the direction angle. Among them, if That is, the first direction angle of the complaint information a i is located in The corresponding cell, the complaint information a i is associated with The corresponding cell; if That is, the first direction angle of the complaint information a i is located in The corresponding cell, the complaint information a i is associated with The corresponding cell; if That is, the first direction angle of the complaint information a i is located in The corresponding cell, the complaint information a i is associated with The corresponding cell.

[0068] In this way, by the above method, multiple complaint information A = [a1, a2, …, a n is associated with the community, and the associated community set can be expressed as C = [c1, c2, …, c n . After obtaining the associated community set, determine the communities in the community set corresponding to multiple deteriorated communities, that is, at least one first community can be obtained. Specifically, take the intersection of the set of multiple deteriorated communities and the associated community set to obtain the set of at least one first community, which can be expressed as S prioi = C ∩ S, and then set at least one first community as the first maintenance priority to realize the confirmation of the maintenance priority of the first community.

[0069] In one embodiment, before calculating the distance between the location where each complaint information is located and the multiple base stations based on the first location information and the second location information, the method further includes:

[0070] Obtain the first time of the multiple complaint information and the second time of each deteriorated community in the multiple deteriorated communities. The first time is the time when the complaint information is sent, and the second time is the time when the index of the community deteriorates and reaches a preset time period;

[0071] Delete the complaint information in the multiple complaint information whose first time is earlier than the target second time to obtain multiple intermediate complaint information. The target second time is the earliest time among the second times of the multiple deteriorated communities;

[0072] The calculating the distance between the location where each complaint information is located and the multiple base stations based on the first location information and the second location information includes:

[0073] Based on the first location information and the second location information, calculate the distance between the location where each intermediate complaint information in the multiple intermediate complaint information is located and the multiple base stations.

[0074] It should be noted that there is a situation where the index of the community deteriorates for a short time and then automatically recovers. This situation will not affect the user experience and does not need to be processed. However, due to the lag and delay of user complaints, there may be a situation where users file complaints when the community deteriorates for a short time. Such complaints cannot accurately reflect that the community needs maintenance, and this part of the complaint information needs to be filtered to improve the accuracy of the complaint information.

[0075] In an embodiment of the present invention, the first time to obtain the multiple complaint information and the second time of each of the multiple deteriorated cells are obtained. The first time is the time when the complaint information is sent, and the second time is the time when the cell index deterioration time reaches a preset time period. Delete the complaint information in the multiple complaint information whose first time is earlier than the target second time to obtain multiple intermediate complaint information. The target second time is the earliest time among the second times of the multiple deteriorated cells. In this way, the complaint information is filtered by the target second time and the first time of the complaint information, so that the filtered complaint information can more accurately reflect whether the deteriorated cell needs maintenance.

[0076] Among them, the second time is the time when the cell index deterioration time reaches a preset time period. After the cell index deterioration time reaches the preset time period, the cell generates a quality difference work order that needs to be processed by maintenance personnel, and the second time is the time when the quality difference work order is generated. In some embodiments, the preset time period is set to T, and the current time is t1. Then the time period for obtaining the identifiers of multiple deteriorated cells can be set to (t1 - T, t1], and t1 - T is the target second time. At this time, the set of multiple deteriorated cells can be expressed as S = [s1, s2, …, s k , and the set of filtered complaint information is expressed as A = [a1, a2, …, a n , where the first time t corresponding to the complaint information a i satisfies t ∈ (t1 - T, t1].

[0077] Specifically, as Figure 3 shown, after obtaining multiple complaint information of the user, determine the target second time corresponding to multiple deteriorated cells, filter the multiple complaint information by the target second time to obtain the set of filtered complaint information A; calculate the associated base station closest to the location where the complaint information is determined, determine the associated cell according to the first direction angle, and repeat the calculation until all the complaint information in the complaint information set A completes the ECI association of the first cell; finally, set the associated first cell as the first maintenance priority.

[0078] In one embodiment, calculating the second maintenance priority of each second cell based on the evaluation information includes:

[0079] Calculating the complaint probability of each second cell based on the historical data; and / or calculating the deterioration score of each second cell based on the deterioration index;

[0080] Determining the second maintenance priority of each second cell based on the complaint probability and / or the deterioration score.

[0081] In an embodiment of the present invention, the complaint probability of each second cell is calculated based on historical data; and / or, the degradation score of each second cell is calculated based on the degradation index; so as to determine the second maintenance priority of each second cell based on the complaint probability and / or the degradation score.

[0082] In some embodiments, when the evaluation information only includes historical data or degradation indicators, the complaint probability or degradation score can be directly set as the value of the second maintenance priority; when the evaluation information only includes historical data and degradation indicators, the weighted sum of the complaint probability and the degradation score is set as the value of the second maintenance priority, so that maintenance personnel can determine the order of maintenance of different second cells through the value of the second maintenance priority.

[0083] In one embodiment, the calculating the complaint probability of each second cell based on the historical data includes:

[0084] Extracting features from the historical data to obtain a feature vector;

[0085] Calculate a first intermediate probability corresponding to the feature vector based on a preset function model;

[0086] Calculating an expression vector corresponding to the feature vector in a preset feature space, then calculating a Mahalanobis distance between the expression vector and a preset mean vector, and converting the Mahalanobis distance into a second intermediate probability;

[0087] The first intermediate probability and the second intermediate probability are weighted to obtain the complaint probability of each second cell.

[0088] It should be noted that due to the subjectivity, lag, inertia, uncertainty and other characteristics of user complaints, there are cases where the network indicators of a cell have deteriorated, affecting the user's network perception, but the user has not filed a complaint, or the delayed complaint has resulted in the failure to associate with the cell. Therefore, it is necessary to conduct complaint prediction analysis on deteriorated cells where indicators have deteriorated but no complaints have occurred, to explore the potential for complaints and thereby improve the user experience.

[0089] In some implementations, the probability corresponding to the feature vector calculated by the preset function model is directly used as the complaint probability. However, since the original data used to train the preset function model may also be inaccurate due to subjective factors of the user, it is specifically manifested in that the degradation of the cell indicators has seriously affected the user's perception, but the user has not reported a complaint to the cell for personal reasons, which leads to inaccurate original data used for model training, resulting in a low accuracy rate of the probability calculated by the preset function model as the complaint probability.

[0090] In an embodiment of the present invention, feature extraction is performed on the historical data to obtain a feature vector; a first intermediate probability corresponding to the feature vector is calculated based on a preset function model; an expression vector corresponding to the feature vector in a preset feature space is calculated, and then the Mahalanobis distance between the expression vector and a preset mean vector is calculated, and the Mahalanobis distance is converted into a second intermediate probability; the first intermediate probability and the second intermediate probability are weighted to obtain the complaint probability of each second cell. In this way, the first intermediate probability is corrected by the second intermediate probability, and compared with directly calculating the complaint probability through the preset function model, the accuracy of the calculated complaint probability can be effectively improved.

[0091] Specifically, as Figure 4 shown, after obtaining the historical data, feature extraction is performed on the historical data to obtain a feature vector; then the feature vector is input into a complaint prediction module and a complaint correction module respectively. The complaint prediction module calculates a first intermediate probability corresponding to the feature vector through a preset function model, and the complaint correction module calculates a second intermediate probability, and then the first intermediate probability and the second intermediate probability are weighted to obtain the complaint probability.

[0092] In some embodiments, after obtaining the historical data, it is necessary to first perform preprocessing such as sorting and cleaning on the historical data to filter out data with a large number of errors and missing values. Among them, if the missing ratio of the data exceeds a set ratio threshold (for example, the set ratio threshold is 80%), the data is deleted; if the missing ratio of the data does not exceed the set ratio threshold, the missing part is filled by linear interpolation. The field format of the historical data after preprocessing can be as shown in the following table.

[0093]

[0094]

[0095] After preprocessing the historical data, feature extraction is performed on the historical data to obtain a feature vector.

[0096] In some embodiments, feature extraction can be performed through a convolutional neural network. Among them, the convolutional neural network extracts features through three fully connected layers, and a batch normalization layer is added after each fully connected layer, and then the linear rectifier function (such as Leaky Relu) is used as the activation function of the network. Specifically, the parameter of the linear rectifier function can be set to 0.01, and the parameter of the dropout of the network can be set to 0.6. The feature vector is extracted through this convolutional neural network model. For example, the input historical data is a one-dimensional array of 1*36, and a feature vector of 1*1024 is obtained through the above convolutional neural network model, denoted as x feat 。

[0097] Among them, the convolutional neural network model can be expressed as: X is historical data, and x feat is a feature vector.

[0098] In some embodiments, the complaint prediction module calculates a first intermediate probability corresponding to the feature vector through a preset function model. It can map the feature vector to a floating-point number through a fully connected layer, and then calculate the floating-point number through the preset function model to obtain the first intermediate probability. The specific process is represented by the following formula:

[0099]

[0100] In the formula is the first intermediate probability, θ T is the parameter of the fully connected layer, and σ is the Sigmoid function model (i.e., the preset function model). The first intermediate probability is obtained through the above formula.

[0101] In one embodiment, the preset function model is obtained in the following manner:

[0102] Obtain second sample data, where the second sample data includes a plurality of positive samples and a plurality of negative samples;

[0103] Train an initial function model based on the plurality of positive samples and the plurality of negative samples to obtain a second intermediate training model;

[0104] Calculate a second loss value corresponding to the second intermediate training model based on the binary cross-entropy loss function;

[0105] In the case where the second loss value is less than a third set loss threshold and the weighted sum of the first loss value and the second loss value is less than a second set loss threshold, set the second intermediate training model as the preset function model.

[0106] Among them, the process of calculating the second loss value Loss pred can be represented by the following formula:

[0107]

[0108] In the above formula, N is the total number of all samples, where the samples with complaints are recorded as positive samples (y (i) = 1), and the samples without complaints are recorded as negative samples (y (i) = 0).

[0109] In some embodiments, the preset feature space is a space composed of the feature vectors of all cells where complaints are sent. The complaint correction module mines the potential probability (i.e., the second intermediate probability) of a complaint occurring in a cell by learning the feature space E of all cells where complaints occur.

[0110] Specifically, in this feature space E, it is assumed that the optimized cells (positive samples) where complaints occur conform to a Gaussian distribution where μ is the mean vector (i.e., the preset mean vector), and Σ is the covariance matrix. The input of this module is the same as that of the complaint prediction module, which is a 1*1024 feature vector x feat , and then a 1*8-dimensional vector is obtained through a fully connected layer. This vector is the expression vector x of this sample in the feature space Ε E , and then through the expression vector x E and the preset mean vector, the Mahalanobis distance is calculated, and then the second intermediate probability is obtained.

[0111] Among them, the calculation process of the second intermediate probability is expressed by the following formula:

[0112]

[0113] In the formula, p(x) is the second intermediate probability, D m (x) is the Mahalanobis Distance. The second intermediate probability is calculated and obtained through the above formula. Among them, p(x) ∈ (0, 1], and the smaller the Mahalanobis distance, the closer the second intermediate probability is to 1.

[0114] In one embodiment, the preset mean vector is obtained in the following manner:

[0115] Obtain the first sample data, and the first sample data includes the sample vectors corresponding to multiple complaint cells where complaints occur;

[0116] Based on the sample data, train the initial Gaussian distribution function to obtain the first intermediate training model;

[0117] Based on the Negative Log-Likelihood (NLL) function, calculate the first loss value corresponding to the first intermediate training model;

[0118] When the first loss value is less than the first set loss threshold, and the weighted sum of the first loss value and the second loss value is less than the second set loss threshold, based on the first intermediate training model and the sample vectors of the multiple complaint cells, calculate and obtain the preset mean vector, and the second loss value is the loss value corresponding to the preset function model.

[0119] It should be noted that through the above negative log-likelihood function, all positive samples can conform to the Gaussian distribution, and then the preset mean vector can be calculated. Specifically, the first loss value Loss mod The calculation process can be expressed by the following formula:

[0120]

[0121]

[0122] where M is the number of all positive samples, is the expression vector of the i-th positive sample in the feature space. The first intermediate model trained by the above formula can calculate the feature vector to obtain the corresponding expression vector of the feature vector.

[0123] In some embodiments, by calculating the overall loss value and constraining the first loss value and the second loss value simultaneously, the accuracy of the preset function model and the Gaussian distribution function model obtained by training is improved.

[0124] Among them, the weighted sum of the first loss value and the second loss value being less than the second set loss threshold can be expressed by the following formula:

[0125] L = Loss pred + λLoss mod ;

[0126] where λ is a weight parameter, and its value range is (0, 1).

[0127] In one embodiment, the deterioration index includes at least one of a cell satisfaction score parameter, a cell application scenario parameter, a cell traffic volume parameter, a cell traffic parameter, a current index parameter, and multiple index parameters in a set time period;

[0128] Calculating the deterioration score of each second cell based on the deterioration index includes:

[0129] Calculating a first factor based on the cell satisfaction score parameter; and / or calculating a second factor based on the cell application scenario parameter; and / or calculating a third factor based on the cell traffic volume parameter and / or the cell traffic parameter; and / or calculating a fourth factor based on the current index parameter; and / or calculating a fifth factor based on the multiple index parameters in the set time period, where the fourth factor is used to characterize the deterioration degree of the current index parameter, and the fifth factor is used to characterize the fluctuation of the multiple index parameters;

[0130] Weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the deterioration score.

[0131] In an embodiment of the present invention, a first factor is calculated based on the cell satisfaction rating parameter; and / or, a second factor is calculated based on the cell application scenario parameter; and / or, a third factor is calculated based on the cell traffic volume parameter and / or the cell traffic parameter; and / or, a fourth factor is calculated based on the current metric parameter; and / or, a fifth factor is calculated based on multiple metric parameters in a set time period, where the fourth factor is used to characterize the degree of deterioration of the current metric parameter, and the fifth factor is used to characterize the fluctuation of multiple metric parameters; the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor are weighted to obtain a deterioration score. In this way, the deterioration score is calculated based on at least one of the cell satisfaction rating parameter, the cell application scenario parameter, the cell traffic volume parameter, the cell traffic parameter, the current metric parameter, and multiple metric parameters in a set time period, so that the deterioration score can reflect the situation of the cell from at least one aspect of the cell satisfaction rating parameter, the cell application scenario parameter, the cell traffic volume parameter, the cell traffic parameter, the current metric parameter, and multiple metric parameters in a set time period, thereby improving the accuracy of determining the second maintenance priority of the cell through the deterioration index.

[0132] In some embodiments, the cell satisfaction rating parameter includes multiple satisfaction scores, and calculating the first factor based on the cell satisfaction rating parameter includes:

[0133] Setting the average value of the multiple satisfaction scores as the first factor.

[0134] It should be noted that the above first factor is used to characterize the user's satisfaction with the cell. By collecting the overall evaluation of the user's network usage experience, the user satisfaction score is associated with the cell where the user is located, and the average value of the user satisfaction scores within the cell is calculated to obtain the first factor for characterizing the user's satisfaction with the cell.

[0135] Among them, the user satisfaction score is usually between [0, 10]. The lower the score, the worse the user's network usage experience, and the more the cell where the user is located should be focused on and maintained.

[0136] The calculation process of the above first factor G(x) can be expressed by the following formula:

[0137]

[0138] In the formula, x is the average value of multiple satisfaction scores.

[0139] In some embodiments, calculating the second factor based on the cell application scenario parameter includes:

[0140] When the cell application scenario parameter is the parameter of the target scenario, setting the second factor as a preset value;

[0141] In the case where the cell application scenario parameter is not the parameter of the target scenario, determine one of the coverage area, coverage scenario, and network mode corresponding to the cell application scenario parameter;

[0142] Query the first score corresponding to the coverage area, the second score corresponding to the coverage scenario, and / or the third score corresponding to the network mode based on a preset scoring table;

[0143] Set the weighted sum of the first score, the second score, and / or the third score as the second factor.

[0144] The above second factor is used to characterize the importance of the cell application scenario. It should be noted that the importance of different scenarios applied to different cells is different. The cell scenario level is an important factor in measuring the cell value. The impact caused by a cell with a high-level scenario when a network quality problem occurs is greater than that of a cell with a general-level scenario. Therefore, it is necessary to consider from the cell level dimension to obtain the second factor used to characterize the importance of the cell application scenario.

[0145] Among them, the preset scoring table can be as shown in the following table:

[0146]

[0147] Through the above table, the first score, the second score, and the third score corresponding to the cell can be obtained.

[0148] Furthermore, for a cell in the target scenario, its importance is much higher than that of other cells. Such cells do not need to calculate the first score, the second score, and the third score, but set the second factor to a preset value to characterize it as the cell with the highest importance level. For example, the target scenarios are hospitals, transportation hubs, etc.

[0149] Among them, the calculation process of the second factor H(x) can be expressed by the following formula:

[0150]

[0151] In the formula, h1(x) is the first score, h2(x) is the second score, and h3(x) is the third score.

[0152] In some embodiments, the calculating the third factor based on the cell traffic parameter and / or the cell traffic volume parameter includes:

[0153] Determine the fourth score corresponding to the cell traffic parameter based on a preset traffic volume level table;

[0154] Determine the fifth score corresponding to the cell traffic volume parameter based on a preset traffic volume level table;

[0155] Set the weighted sum of the fourth score and the fifth score as the third factor.

[0156] It should be noted that the cell traffic volume and traffic size are important indicators reflecting the importance of the cell. When a cell with a large traffic volume and traffic size deteriorates, the resulting service loss and impact on customer perception are also relatively large. Therefore, the deterioration score is evaluated from the cell traffic volume parameter and the traffic parameter.

[0157] Specifically, according to the characteristics of the cell traffic volume and traffic, the cell traffic volume and traffic can be classified to obtain the corresponding fourth score or fifth score.

[0158] For example, when the cell traffic volume parameter is less than 100 Erl, it is defined as a low-traffic cell; when the cell traffic volume parameter is greater than 200 Erl, it is defined as a high-traffic cell. At this time, the calculation process of the fourth score i1(x) can be expressed by the following formula:

[0159]

[0160] In the formula, x is the cell traffic volume parameter.

[0161] Also, for example, when the cell traffic parameter is less than 50 GB, it is defined as a low-traffic cell; when the cell traffic parameter is greater than 250 GB, it is defined as a high-traffic cell. At this time, the calculation process of the fifth score i2(y) can be expressed by the following formula:

[0162]

[0163] In the formula, y is the cell traffic parameter.

[0164] After calculating the fourth score and the fifth score, the third factor I(x, y) can be calculated through the following formula:

[0165] I(x, y) = i1(x) + i2(y), I(x, y) ∈ (0, 1];

[0166] The third factor is calculated through the above formula.

[0167] In one embodiment, calculating the fourth factor based on the current indicator parameter includes:

[0168] Obtain the deterioration threshold corresponding to the current indicator parameter;

[0169] When the current indicator parameter is a first-type parameter, set the quotient of the first difference and the second difference as the intermediate score. The first difference is the difference between 1 and the current indicator parameter, and the second difference is the difference between 1 and the deterioration threshold. The first-type parameter is a parameter for which deterioration occurs when the value of the indicator parameter decreases;

[0170] When the current metric parameter is a second type of parameter, the quotient of the current metric parameter and the deterioration threshold is the intermediate score, and the second type of parameter is a parameter whose deterioration occurs as the value of the metric parameter increases;

[0171] Normalize the intermediate score to obtain the fourth factor.

[0172] It should be noted that the degree of deterioration of the cell's metrics is also an important factor in reflecting whether a cell with poor quality needs to be preferentially optimized. The more severely deteriorated the cell's metrics are, the more it should be preferentially maintained. Among them, due to the different natures of the metrics, some metrics are better when they are larger, and some metrics are better when they are smaller, making it difficult to unify the measurement. Therefore, when constructing the function of the degree of deterioration of the metrics, it is necessary to first determine whether the current preparation parameter belongs to the first type of parameter or the second type of parameter, and then calculate to obtain the fourth factor.

[0173] Among them, the first type of parameter can be handover success rate, radio access success rate, RRC connection establishment success rate, Flow establishment success rate, gNB inter-handover success rate, gNB intra-handover success rate, NR to LTE handover-based EPSFB success rate, QoS Flow establishment success rate per slice, NG interface UE-related logical signaling connection establishment success rate, etc.

[0174] The second type of parameter can be radio link failure rate, Flow disconnection rate, RRC connection reestablishment ratio, average utilization rate of uplink / downlink PRBs, downlink packet loss rate of the cell's RLC layer, MAC layer uplink / downlink block error rate, uplink / downlink HARQ retransmission ratio, high single-stream ratio, uplink packet loss rate of the cell's PDCP layer, downlink packet loss rate of the cell's RLC layer, VoNR service Flow disconnection rate, ViNR service Flow disconnection rate, etc.

[0175] The calculation process of the above intermediate score j(x) can be expressed by the following formula:

[0176]

[0177] The intermediate score is calculated through the above formula. The larger the intermediate score, the more severe the deterioration degree of the current metric parameter.

[0178] Furthermore, it is also necessary to normalize the intermediate score to obtain the fourth factor. The calculation process is shown in the following formula:

[0179]

[0180] In the formula, J(x) is the fourth factor, j min (x) is the preset minimum intermediate score, j max (x) is the preset maximum intermediate score, and J(x) ∈ [0, 1].

[0181] In one embodiment, calculating the fifth factor based on multiple metric parameters in the set time period includes:

[0182] Calculating the average value and standard deviation of the multiple metric parameters;

[0183] Setting the quotient of the standard deviation and the average value as the coefficient of variation;

[0184] Normalizing the coefficient of variation to obtain the fifth factor.

[0185] It should be noted that metric parameters may, due to reasons such as sudden interference, human factors, or temporary equipment failures, cause the metrics to deteriorate repeatedly, recover, deteriorate again, and finally recover automatically. Its characteristic is that the metrics show large fluctuations within the deterioration cycle. For the deterioration of metrics caused by certain specific failures or unreasonable parameter configurations, the deterioration trend of the metrics shows the characteristic of continuous deterioration. Therefore, the volatility of cell metrics is also an important reference for calculating the deterioration score of the cell. The detection of metric volatility can be characterized by calculating the coefficient of variation of the metrics within the deterioration cycle.

[0186] Among them, the coefficient of variation is an index used to measure the relative degree of data fluctuation, and is the ratio of the standard deviation to the average value. The larger the coefficient of variation, the greater the relative degree of data fluctuation. Suppose the number of metric parameters of the current metric parameter x within the deterioration cycle is n, which are (x1, x2,..., x n ), then the average value μ(x) of the current metric parameter x can be calculated by the following formula:

[0187]

[0188] The standard deviation σ(x) can be calculated by the following formula:

[0189]

[0190] The coefficient of variation CV(x) can be calculated by the following formula:

[0191]

[0192] Normalizing the coefficient of variation to obtain the fifth factor k(x), which can be expressed by the following formula:

[0193]

[0194] In the formula, CV min (x) is the preset minimum coefficient of variation, and CV max (x) is the preset maximum coefficient of variation.

[0195] Furthermore, since the higher the coefficient of variation of an indicator, the greater the volatility of the indicator, and the higher the probability that the indicator can recover by itself under the influence of sudden factors, the lower the priority required for the poor-quality cell. Therefore, the fifth factor K(x) of the cell can also be expressed as:

[0196] K(x) = 1 - k(x).

[0197] In one embodiment, before weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the degradation score, it further includes:

[0198] Determining a judgment matrix, where the judgment matrix is used to characterize the relative importance between any two factors among the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor;

[0199] Based on the judgment matrix, assigning a first coefficient to the first factor, a second coefficient to the second factor, a third coefficient to the third factor, a fourth coefficient to the fourth factor, and / or a fifth coefficient to the fifth factor;

[0200] The step of weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the degradation score includes:

[0201] Weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor based on the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, and / or the fifth coefficient to obtain the degradation score.

[0202] The above judgment matrix is a matrix obtained based on the analytic hierarchy process. The analytic hierarchy process is a decision-making method that decomposes the elements related to the decision into levels such as goals, criteria, and solutions, and then conducts qualitative and quantitative analysis on this basis. The analytic hierarchy process provides a new, simple, and practical decision-making method for studying complex systems composed of many interrelated and mutually restricted factors. Therefore, in the present invention, the weight calculation of different factors can be performed through the analytic hierarchy process.

[0203] Among them, the objectives of the decision-making, the considered factors (decision-making criteria), and the decision-making objects are divided into the highest level, the middle level, and the lowest level according to the mutual relationships among different factors. Among them, the highest level is the weight coefficients of different factors, the middle level is divided into influencing factors such as the probability of complaints occurring in the community, the community satisfaction score, the community scene level, the community traffic volume, the traffic, the degree of index deterioration, and the degree of index fluctuation, and the lowest level is the value score situation of the poor-quality communities in the same batch. The judgment matrix of the analytic hierarchy process adopts a relative scale to minimize the difficulty of comparing many factors with different natures as much as possible, thereby improving the accuracy. For example, for n evaluation indicators, by using the 1-9 scale method, that is, the binary relative comparison method, the relative importance between each indicator is judged, and the matrix formed according to the pairwise comparison results, that is, the judgment matrix X, is used to determine the relative importance among n sub-factors at a certain level.

[0204] Specifically, the judgment matrix can be expressed by the following formula:

[0205]

[0206] Among them, x ij represents the importance degree of factor i relative to factor j. The judgment matrix scale and its meanings are shown in the following table:

[0207]

[0208]

[0209] Through the above table, a judgment matrix X is constructed around the complaint probability, the first factor, the second factor, the third factor, the fourth factor, and the fifth factor, as specifically shown below

[0210]

[0211] After obtaining the judgment matrix, based on the judgment matrix, a first coefficient is assigned to the first factor, a second coefficient is assigned to the second factor, a third coefficient is assigned to the third factor, a fourth coefficient is assigned to the fourth factor, and / or a fifth coefficient is assigned to the fifth factor, so as to assign weight coefficients W i , and a weight coefficient set W = (W1, W2,..., W n ) T is formed, and

[0212] In some embodiments, the weight coefficients can be calculated by the arithmetic mean method. Specifically, it is expressed by the following formula:

[0213]

[0214] Substitute the judgment matrix X into the weight calculation formula to obtain the weight coefficient set corresponding to different factors:

[0215] W = (0.43, 0.15, 0.23, 0.06, 0.09, 0.03) T 。

[0216] In some embodiments, to ensure the rationality of the conclusions obtained by applying the analytic hierarchy process, it is also necessary to test the consistency of the judgment matrix. According to the conclusions of matrix theory, when the judgment matrix cannot guarantee complete consistency, the eigenvalues of the corresponding judgment matrix will also change. Therefore, the change of the eigenvalues of the judgment matrix can be used to test the degree of judgment consistency.

[0217] Specifically, introducing the consistency index of the judgment matrix in the analytic hierarchy process is used to check the consistency of judgment thinking. The index is denoted as CI. The larger the CI value, the more severely the judgment matrix deviates from complete consistency, and the smaller the CI value, the closer the judgment matrix is to complete consistency.

[0218] Among them, CI can be calculated by the following formula:

[0219] CI = (λ max - n) / (n - 1);

[0220] In the formula, λ max is the largest eigenvalue of the judgment matrix, which is specifically calculated by the following formula:

[0221]

[0222] To measure the magnitude of CI, the random consistency index RI is introduced. RI is related to the order of the judgment matrix. Generally, the larger the matrix order, the greater the possibility of random deviation from consistency. The corresponding relationship of the random consistency index is as follows:

[0223]

[0224] Considering that the deviation from consistency may be caused by random reasons, it is necessary to compare CI with the random consistency index RI to obtain the test coefficient CR.

[0225]

[0226] Among them, when CR < 0.1, it is considered that the consistency of the judgment matrix is acceptable.

[0227] Exemplarily, substituting the weight matrix W into the eigenvalue calculation formula, we can obtain λ max = 6.48, CI = 0.096, RI = 1.26, CR = 0.08 < 0.1, which meets the matrix consistency judgment.

[0228] In some embodiments, the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor are weighted based on the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, and / or the fifth coefficient to obtain the degradation score F(x), which can be represented by the following formula:

[0229] F(x) = w2G(x) + w3H(x) + w4I(x) + w5J(x) + w6K(x).

[0230] Further, the complaint probability P(x) can also be assigned a weight coefficient in the above manner, and then weighted with the degradation score to obtain the numerical value F(x) of the final second maintenance priority, which is specifically represented by the following formula:

[0231] F(x) = w1P(x) + w2G(x) + w3H(x) + w4I(x) + w5J(x) + w6K(x).

[0232] Please refer to Figure 5 , Figure 5 which is the structural diagram of a cell maintenance priority determination device provided by an embodiment of the present invention. As Figure 5 shown, the cell maintenance priority determination device 500 includes:

[0233] A first acquisition module 501, configured to acquire a plurality of complaint information and identifiers of a plurality of degraded cells;

[0234] A determination module 502, configured to determine at least one first cell associated with the plurality of complaint information, and set the at least one first cell as the first maintenance priority, where the at least one first cell is a cell among the plurality of degraded cells;

[0235] A second acquisition module 503, configured to acquire evaluation information of each second cell in a plurality of second cells, where the evaluation information includes a degradation index and / or historical data corresponding to the degradation index, and the plurality of second cells are cells other than the at least one first cell among the plurality of degraded cells;

[0236] A calculation module 504, configured to calculate the second maintenance priority of each second cell based on the evaluation information, where the second maintenance priority is lower than the first maintenance priority.

[0237] In one embodiment, each piece of complaint information in the plurality of complaint information includes first location information, and the determination module 502 includes:

[0238] A first acquisition sub-module, configured to acquire second location information corresponding to a plurality of base stations, where each base station in the plurality of base stations includes at least one degraded cell, and different degraded cells correspond to different azimuth angle ranges of the base stations;

[0239] A first calculation sub-module, configured to calculate the distances between the locations where each complaint information is located and the multiple base stations based on the first location information and the second location information;

[0240] A first association sub-module, configured to set the base station with the minimum distance corresponding to the target complaint information as the associated base station of the target complaint information, where the target complaint information is one of the multiple complaint information;

[0241] A second calculation sub-module, configured to calculate a first direction angle between the location where the target complaint information is located and the associated base station;

[0242] A second association sub-module, configured to determine a first cell associated with the target complaint information based on the first direction angle, where the first cell is a cell included in the associated base station, and the direction angle range corresponding to the first cell includes the first direction angle.

[0243] In one embodiment, the determination module 502 further includes:

[0244] A second acquisition sub-module, configured to acquire a first time of the multiple complaint information and a second time of each of the multiple degraded cells, where the first time is the time when the complaint information is sent, and the second time is the time when the index of the cell degrades to reach a preset time period;

[0245] A deletion sub-module, configured to delete the complaint information among the multiple complaint information whose first time is earlier than the target second time to obtain multiple intermediate complaint information, where the target second time is the earliest time among the second times of the multiple degraded cells;

[0246] The first calculation sub-module includes:

[0247] A first calculation unit, configured to calculate the distances between the locations where each of the multiple intermediate complaint information is located and the multiple base stations based on the first location information and the second location information.

[0248] In one embodiment, the calculation module 504 includes:

[0249] A third calculation sub-module, configured to calculate a complaint probability of each second cell based on the historical data; and / or calculate a degradation score of each second cell based on the degradation index;

[0250] A determination sub-module, configured to determine a second maintenance priority of each second cell based on the complaint probability and / or the degradation score.

[0251] In one embodiment, the third calculation sub-module includes:

[0252] An extraction unit for extracting features from the historical data to obtain a feature vector;

[0253] A second calculation unit for calculating a first intermediate probability corresponding to the feature vector based on a preset function model;

[0254] A third calculation unit for calculating an expression vector corresponding to the feature vector in a preset feature space, then calculating a Mahalanobis distance between the expression vector and a preset mean vector, and converting the Mahalanobis distance into a second intermediate probability;

[0255] A first weighting unit for weighting the first intermediate probability and the second intermediate probability to obtain the complaint probability of each second cell.

[0256] In one embodiment, the preset mean vector is obtained by the following method:

[0257] Obtain first sample data, where the first sample data includes sample vectors corresponding to multiple complaint cells where complaints occur;

[0258] Train an initial Gaussian distribution function based on the sample data to obtain a first intermediate training model;

[0259] Calculate a first loss value corresponding to the first intermediate training model based on a negative log-likelihood function;

[0260] In the case where the first loss value is less than a first set loss threshold and the weighted sum of the first loss value and a second loss value is less than a second set loss threshold, calculate the preset mean vector based on the first intermediate training model and the sample vectors of the multiple complaint cells, and the second loss value is the loss value corresponding to the preset function model.

[0261] In one embodiment, the preset function model is obtained by the following method:

[0262] Obtain second sample data, where the second sample data includes multiple positive samples and multiple negative samples;

[0263] Train an initial function model based on the multiple positive samples and the multiple negative samples to obtain a second intermediate training model;

[0264] Calculate a second loss value corresponding to the second intermediate training model based on a binary cross-entropy loss function;

[0265] When the second loss value is less than the third set loss threshold and the weighted sum of the first loss value and the second loss value is less than the second set loss threshold, set the second intermediate training model as the preset function model.

[0266] In one embodiment, the deterioration index includes at least one of a cell satisfaction score parameter, a cell application scenario parameter, a cell traffic volume parameter, a cell traffic parameter, a current index parameter, and multiple index parameters in a set time period.

[0267] The third calculation sub-module includes:

[0268] A fourth calculation unit, configured to calculate a first factor based on the cell satisfaction score parameter; and / or calculate a second factor based on the cell application scenario parameter; and / or calculate a third factor based on the cell traffic volume parameter and / or the cell traffic parameter; and / or calculate a fourth factor based on the current index parameter; and / or calculate a fifth factor based on the multiple index parameters in the set time period, where the fourth factor is used to characterize the deterioration degree of the current index parameter, and the fifth factor is used to characterize the fluctuation of the multiple index parameters.

[0269] A second weighting unit, configured to weight the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the deterioration score.

[0270] In one embodiment, the fourth calculation unit includes:

[0271] An acquisition subunit, configured to acquire a deterioration threshold corresponding to the current index parameter.

[0272] A first setting subunit, configured to, when the current index parameter is a first type of parameter, set the quotient of a first difference and a second difference as an intermediate score, where the first difference is the difference between 1 and the current index parameter, the second difference is the difference between 1 and the deterioration threshold, and the first type of parameter is a parameter for which deterioration occurs when the value of the index parameter decreases.

[0273] A second setting subunit, configured to, when the current index parameter is a second type of parameter, set the quotient of the current index parameter and the deterioration threshold as the intermediate score, where the second type of parameter is a parameter for which deterioration occurs when the value of the index parameter increases.

[0274] A first normalization subunit, configured to perform normalization processing on the intermediate score to obtain the fourth factor.

[0275] In one embodiment, the fourth calculation unit includes:

[0276] A first calculation sub-unit for calculating the average value and standard deviation of the multiple index parameters;

[0277] A second calculation sub-unit for setting the quotient of the standard deviation and the average value as the coefficient of variation;

[0278] A normalization sub-unit for normalizing the coefficient of variation to obtain the fifth factor.

[0279] In one embodiment, the third calculation sub-module further includes:

[0280] A determination unit for determining a judgment matrix, where the judgment matrix is used to characterize the relative importance between any two of the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor;

[0281] An allocation unit for allocating a first coefficient to the first factor, a second coefficient to the second factor, a third coefficient to the third factor, a fourth coefficient to the fourth factor, and / or a fifth coefficient to the fifth factor based on the judgment matrix;

[0282] The second weighting unit includes:

[0283] A weighting sub-unit for weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor based on the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, and / or the fifth coefficient to obtain the deterioration score.

[0284] The cell maintenance priority determination device provided by the embodiments of the present invention can implement each process of the above-mentioned cell maintenance priority determination method in various embodiments, with the technical features corresponding one by one and achieving the same technical effects. To avoid repetition, it will not be elaborated here.

[0285] It should be noted that the cell maintenance priority determination device in the embodiments of the present invention can be a device, or a component, an integrated circuit, or a chip in an electronic device.

[0286] The embodiments of the present invention further provide an electronic device, including: a processor, a memory, and a program stored on the memory and executable on the processor. When the program is executed by the processor, it implements each process of the above-mentioned cell maintenance priority determination method embodiments and can achieve the same technical effects. To avoid repetition, it will not be elaborated here.

[0287] Specifically, as shown in Figure 6 The embodiments of the present invention further provide an electronic device, including a bus 601, a transceiver 602, an antenna 603, a bus interface 604, a processor 605, and a memory 606.

[0288] The transceiver 602 is configured to obtain a plurality of complaint information and identifiers of a plurality of deteriorated cells;

[0289] The processor 605 is configured to determine at least one first cell associated with the plurality of complaint information, and set the at least one first cell to a first maintenance priority, where the at least one first cell is a cell among the plurality of deteriorated cells;

[0290] The transceiver 602 is further configured to obtain evaluation information of each second cell among a plurality of second cells, where the evaluation information includes a deterioration index and / or historical data corresponding to the deterioration index, and the plurality of second cells are cells among the plurality of deteriorated cells other than the at least one first cell;

[0291] The processor 605 is further configured to calculate a second maintenance priority of each second cell based on the evaluation information, where the second maintenance priority is lower than the first maintenance priority.

[0292] In one embodiment, each complaint information among the plurality of complaint information includes first location information, and determining at least one first cell associated with the plurality of complaint information includes:

[0293] Obtain second location information corresponding to a plurality of base stations, where each base station among the plurality of base stations includes at least one deteriorated cell, and different deteriorated cells correspond to different azimuth ranges of the base stations;

[0294] Based on the first location information and the second location information, calculate the distances between the location where each complaint information is located and the plurality of base stations;

[0295] Set the base station with the minimum distance corresponding to the target complaint information as the associated base station of the target complaint information, where the target complaint information is one of the plurality of complaint information;

[0296] Calculate a first azimuth between the location where the target complaint information is located and the associated base station;

[0297] Based on the first azimuth, determine a first cell associated with the target complaint information, where the first cell is a cell included in the associated base station, and the azimuth range corresponding to the first cell includes the first azimuth.

[0298] In one embodiment, the transceiver 602 is further configured to obtain a first time of the plurality of complaint information, and a second time of each deteriorated cell among the plurality of deteriorated cells, where the first time is the time when the complaint information is sent, and the second time is the time when the index of the cell deteriorates and reaches a preset time period;

[0299] The processor 605 is further configured to delete the complaint information in the multiple pieces of complaint information whose first time is earlier than the target second time, so as to obtain multiple intermediate complaint information, where the target second time is the earliest time among the second times of the multiple deteriorated cells;

[0300] Calculating the distance between the location where each complaint information is located and the multiple base stations based on the first location information and the second location information includes:

[0301] Based on the first location information and the second location information, calculate the distance between the location where each intermediate complaint information in the multiple intermediate complaint information is located and the multiple base stations.

[0302] In one embodiment, calculating the second maintenance priority of each second cell based on the evaluation information includes:

[0303] Calculating the complaint probability of each second cell based on the historical data; and / or calculating the deterioration score of each second cell based on the deterioration index;

[0304] Determine the second maintenance priority of each second cell based on the complaint probability and / or the deterioration score.

[0305] In one embodiment, calculating the complaint probability of each second cell based on the historical data includes:

[0306] Performing feature extraction on the historical data to obtain a feature vector;

[0307] Calculating a first intermediate probability corresponding to the feature vector based on a preset function model;

[0308] Calculating an expression vector corresponding to the feature vector in a preset feature space, then calculating the Mahalanobis distance between the expression vector and a preset mean vector, and converting the Mahalanobis distance into a second intermediate probability;

[0309] Weight the first intermediate probability and the second intermediate probability to obtain the complaint probability of each second cell.

[0310] In one embodiment, the preset mean vector is obtained in the following manner:

[0311] Obtain first sample data, where the first sample data includes sample vectors corresponding to multiple complaint cells where complaints occur;

[0312] Train an initial Gaussian distribution function based on the sample data to obtain a first intermediate training model;

[0313] Calculate the first loss value corresponding to the first intermediate training model based on the negative log-likelihood function;

[0314] In the case where the first loss value is less than the first set loss threshold and the weighted sum of the first loss value and the second loss value is less than the second set loss threshold, calculate the preset mean vector based on the first intermediate training model and the sample vectors of the multiple complaint cells, where the second loss value is the loss value corresponding to the preset function model.

[0315] In one embodiment, the preset function model is obtained by the following method:

[0316] Obtain second sample data, where the second sample data includes a plurality of positive samples and a plurality of negative samples;

[0317] Train an initial function model based on the plurality of positive samples and the plurality of negative samples to obtain a second intermediate training model;

[0318] Calculate the second loss value corresponding to the second intermediate training model based on the binary cross-entropy loss function;

[0319] In the case where the second loss value is less than the third set loss threshold and the weighted sum of the first loss value and the second loss value is less than the second set loss threshold, set the second intermediate training model as the preset function model.

[0320] In one embodiment, the degradation index includes at least one of a cell satisfaction score parameter, a cell application scenario parameter, a cell traffic volume parameter, a cell traffic parameter, a current index parameter, and a plurality of index parameters in a set time period;

[0321] The calculating the degradation score of each second cell based on the degradation index includes:

[0322] Calculate a first factor based on the cell satisfaction score parameter; and / or calculate a second factor based on the cell application scenario parameter; and / or calculate a third factor based on the cell traffic volume parameter and / or the cell traffic parameter; and / or calculate a fourth factor based on the current index parameter; and / or calculate a fifth factor based on the plurality of index parameters in the set time period, where the fourth factor is used to characterize the degradation degree of the current index parameter, and the fifth factor is used to characterize the fluctuation of the plurality of index parameters;

[0323] Weight the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the degradation score.

[0324] In one embodiment, the calculating the fourth factor based on the current index parameter includes:

[0325] Obtain the deterioration threshold corresponding to the current metric parameter;

[0326] When the current metric parameter is a first type of parameter, set the quotient of the first difference and the second difference as the intermediate score. The first difference is the difference between 1 and the current metric parameter, the second difference is the difference between 1 and the deterioration threshold, and the first type of parameter is a parameter for which deterioration occurs when the value of the metric parameter decreases;

[0327] When the current metric parameter is a second type of parameter, set the quotient of the current metric parameter and the deterioration threshold as the intermediate score. The second type of parameter is a parameter for which deterioration occurs when the value of the metric parameter increases;

[0328] Normalize the intermediate score to obtain the fourth factor.

[0329] In one embodiment, calculating the fifth factor based on multiple metric parameters in the set time period includes:

[0330] Calculate the average value and standard deviation of the multiple metric parameters;

[0331] Set the quotient of the standard deviation and the average value as the coefficient of variation;

[0332] Normalize the coefficient of variation to obtain the fifth factor.

[0333] In one embodiment, before weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the deterioration score, further includes:

[0334] Determine a judgment matrix, where the judgment matrix is used to characterize the relative importance between any two of the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor;

[0335] Based on the judgment matrix, assign a first coefficient to the first factor, a second coefficient to the second factor, a third coefficient to the third factor, a fourth coefficient to the fourth factor, and / or a fifth coefficient to the fifth factor;

[0336] The step of weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the deterioration score includes:

[0337] Based on the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, and / or the fifth coefficient, weight the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the deterioration score.

[0338] In Figure 6 it, the bus architecture (represented by bus 601) may include any number of interconnected buses and bridges. Bus 601 links together various circuits including one or more processors represented by processor 605 and a memory represented by memory 606. Bus 601 may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and thus will not be further described herein. Bus interface 604 provides an interface between bus 601 and transceiver 602. Transceiver 602 may be a single component or multiple components, such as multiple receivers and transmitters, providing units for communicating with various other devices over a transmission medium. Data processed by processor 605 is transmitted over a wireless medium via antenna 603. Further, antenna 603 also receives data and transmits the data to processor 605.

[0339] Processor 605 is responsible for managing bus 601 and general processing, and may also provide various functions including timing, peripheral interface, voltage regulation, power management, and other control functions. Memory 606 may be used to store data used by processor 605 during operation.

[0340] Optionally, processor 605 may be a CPU, ASIC, FPGA, or CPLD.

[0341] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it implements each process of the above-mentioned embodiment of the method for determining the cell maintenance priority and can achieve the same technical effect. To avoid repetition, it will not be elaborated herein. Among them, the computer-readable storage medium is, for example, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disc, etc.

[0342] The present invention also provides a computer program product including computer instructions. When the computer instructions are executed by a processor, they implement each process of the above-mentioned Figure 1 corresponding embodiment of the method for determining the cell maintenance priority and can achieve the same technical effect. To avoid repetition, it will not be elaborated herein.

[0343] It should be noted that in this text, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or device that includes a series of elements not only includes those elements but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or device that includes such element.

[0344] From the description of the above embodiments, those skilled in the art can clearly understand that the above-described embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of the present invention.

[0345] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the purpose of the present invention and the scope protected by the claims, and all of them fall within the protection scope of the present invention.

Claims

1. A method for determining the priority of cell maintenance, characterized in that, Including: Obtain multiple complaint information and the cell unique identifier ECI of multiple degraded cells; Determine the ECI of at least one first cell associated with the multiple complaint information, and set the at least one first cell to the first maintenance priority, where the at least one first cell is a cell among the multiple degraded cells; Obtain the evaluation information of each second cell in multiple second cells, where the evaluation information includes a degradation index and / or historical data corresponding to the degradation index, and the multiple second cells are cells among the multiple degraded cells other than the at least one first cell; Calculate the second maintenance priority of each second cell based on the evaluation information, and the second maintenance priority is lower than the first maintenance priority.

2. The method according to claim 1, wherein, Each complaint information in the multiple complaint information includes first location information. The determining the ECI of at least one first cell associated with the multiple complaint information includes: Obtain the second location information corresponding to multiple base stations, where each base station among the multiple base stations includes at least one degraded cell, and different degraded cells correspond to different angular range of directions of the base stations; Based on the first location information and the second location information, calculate the distance between the location where each complaint information is located and the multiple base stations; Set the base station with the smallest distance corresponding to the target complaint information as the associated base station of the target complaint information, where the target complaint information is one complaint information among the multiple complaint information; Calculate the first direction angle between the location where the target complaint information is located and the associated base station; Based on the first direction angle, determine the ECI of the first cell associated with the target complaint information, where the first cell is a cell included in the associated base station, and the angular range corresponding to the first cell includes the first direction angle.

3. The method according to claim 2, wherein Before the calculating the distance between the location where each complaint information is located and the multiple base stations based on the first location information and the second location information, the method further includes: Obtain the first time of the multiple complaint information and the second time of each degraded cell in the multiple degraded cells, where the first time is the time when the complaint information is sent, and the second time is the time when the degradation time of the cell index reaches a preset time period; Delete the complaint information in the multiple complaint information whose first time is earlier than the target second time to obtain multiple intermediate complaint information, where the target second time is the earliest time among the second times of the multiple degraded cells; The calculating the distance between the location where each complaint information is located and the multiple base stations based on the first location information and the second location information includes: Based on the first location information and the second location information, calculate the distance between the location where each intermediate complaint information in the multiple intermediate complaint information is located and the multiple base stations.

4. The method according to claim 1, wherein The calculating the second maintenance priority of each second cell based on the evaluation information includes: Calculate the complaint probability of each second cell based on the historical data; and / or calculate the degradation score of each second cell based on the degradation index; Determine the second maintenance priority of each second cell based on the complaint probability and / or the degradation score.

5. The method according to claim 4, wherein Calculating the complaint probability of each second cell based on the historical data includes: Performing feature extraction on the historical data to obtain a feature vector; Calculating a first intermediate probability corresponding to the feature vector based on a preset function model; Calculating an expression vector corresponding to the feature vector in a preset feature space, then calculating the Mahalanobis distance between the expression vector and a preset mean vector, and converting the Mahalanobis distance into a second intermediate probability; Weighting the first intermediate probability and the second intermediate probability to obtain the complaint probability of each second cell.

6. The method according to claim 5, wherein The preset mean vector is obtained by the following method: Obtain first sample data, where the first sample data includes sample vectors corresponding to multiple complaint cells where complaints occur; Training an initial Gaussian distribution function based on the sample data to obtain a first intermediate training model; Calculating a first loss value corresponding to the first intermediate training model based on a negative log-likelihood function; When the first loss value is less than a first set loss threshold and the weighted sum of the first loss value and a second loss value is less than a second set loss threshold, based on the first intermediate training model and the sample vectors of the multiple complaint cells, calculate and obtain the preset mean vector, where the second loss value is the loss value corresponding to the preset function model.

7. The method according to claim 5, characterized in that The preset function model is obtained by the following method: Obtain second sample data, where the second sample data includes multiple positive samples and multiple negative samples; Training an initial function model based on the multiple positive samples and the multiple negative samples to obtain a second intermediate training model; Calculating a second loss value corresponding to the second intermediate training model based on a binary cross-entropy loss function; When the second loss value is less than a third set loss threshold and the weighted sum of the first loss value and the second loss value is less than a second set loss threshold, set the second intermediate training model as the preset function model.

8. The method according to claim 4, wherein The degradation index includes at least one of a cell satisfaction score parameter, a cell application scenario parameter, a cell traffic volume parameter, a cell traffic parameter, a current index parameter, and multiple index parameters in a set time period; Calculating the degradation score of each second cell based on the degradation index includes: Calculating a first factor based on the cell satisfaction score parameter; and / or calculating a second factor based on the cell application scenario parameter; and / or calculating a third factor based on the cell traffic volume parameter and / or the cell traffic parameter; and / or calculating a fourth factor based on the current index parameter; and / or calculating a fifth factor based on the multiple index parameters in the set time period, where the fourth factor is used to characterize the degradation degree of the current index parameter, and the fifth factor is used to characterize the fluctuation of the multiple index parameters; Weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the degradation score.

9. The method according to claim 8, wherein Calculating the fourth factor based on the current index parameter includes: Obtain the degradation threshold corresponding to the current index parameter; When the current index parameter is a first - type parameter, set the quotient of the first difference and the second difference as the intermediate score. The first difference is the difference between 1 and the current index parameter, the second difference is the difference between 1 and the deterioration threshold, and the first - type parameter is a parameter for which deterioration occurs when the value of the index parameter decreases; When the current index parameter is a second - type parameter, set the quotient of the current index parameter and the deterioration threshold as the intermediate score. The second - type parameter is a parameter for which deterioration occurs when the value of the index parameter increases; Normalize the intermediate score to obtain the fourth factor.

10. The method according to claim 8, characterized in that, Calculating the fifth factor based on multiple index parameters in the set time period includes: Calculate the average value and standard deviation of the multiple index parameters; Set the quotient of the standard deviation and the average value as the coefficient of variation; Normalize the coefficient of variation to obtain the fifth factor.

11. The method according to claim 8, characterized in that, Before weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the deterioration score, it further includes: Determine a judgment matrix, which is used to characterize the relative importance between any two of the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor; Based on the judgment matrix, assign a first coefficient to the first factor, a second coefficient to the second factor, a third coefficient to the third factor, a fourth coefficient to the fourth factor, and / or a fifth coefficient to the fifth factor; Weighting the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the deterioration score includes: Based on the first coefficient, the second coefficient, the third coefficient, the fourth coefficient, and / or the fifth coefficient, weight the first factor, the second factor, the third factor, the fourth factor, and / or the fifth factor to obtain the deterioration score.

12. A device for determining the priority of cell maintenance, characterized in that, It includes: A first acquisition module, configured to acquire a plurality of complaint information and the cell - unique identifier ECI of a plurality of deteriorated cells; A determination module, configured to determine the ECI of at least one first cell associated with the plurality of complaint information, and set the at least one first cell as the first maintenance priority level. The at least one first cell is a cell among the plurality of deteriorated cells; A second acquisition module, configured to acquire the evaluation information of each second cell in a plurality of second cells. The evaluation information includes a deterioration index and / or historical data corresponding to the deterioration index. The plurality of second cells are cells other than the at least one first cell among the plurality of deteriorated cells; A calculation module, configured to calculate the second maintenance priority level of each second cell based on the evaluation information. The second maintenance priority level is lower than the first maintenance priority level.

13. An electronic device, characterized in that, It includes a transceiver and a processor, The transceiver is configured to acquire a plurality of complaint information and the cell - unique identifier ECI of a plurality of deteriorated cells; The processor is configured to determine the ECI of at least one first cell associated with the multiple complaint messages, and set the at least one first cell as the first maintenance priority level, where the at least one first cell is a cell among the multiple deteriorated cells; The transceiver is further configured to obtain the evaluation information of each second cell among the multiple second cells, where the evaluation information includes a deterioration index and / or historical data corresponding to the deterioration index, and the multiple second cells are cells among the multiple deteriorated cells other than the at least one first cell; The processor is further configured to calculate the second maintenance priority level of each second cell based on the evaluation information, and the second maintenance priority level is lower than the first maintenance priority level.

14. An electronic device, characterized in that, Comprising: A processor, a memory, and a program stored on the memory and executable on the processor, where when the program is executed by the processor, the steps of the cell maintenance priority level determination method according to any one of claims 1 to 11 are implemented.

15. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the cell maintenance priority level determination method according to any one of claims 1 to 11 are implemented.

16. A computer program product, characterized in that, Comprising computer instructions, and when the computer instructions are executed by a processor, the steps of the cell maintenance priority level determination method according to any one of claims 1 to 11 are implemented.

Images