Network complaint root cause positioning method, device and application

Abstract

This application relates to the field of mobile communications, providing a method, apparatus, and application for locating the root cause of network complaints. The method includes: constructing multiple independent trees based on a set of network performance indicators for customer complaints and a set of network performance indicators for non-customer complaints using the isolated forest algorithm; determining whether the set of network performance indicators for customer complaints is abnormal based on the average height of the set of network performance indicators in the multiple independent trees; when the set of network performance indicators for customer complaints is abnormal, determining abnormal indicators based on the path height of the leaf nodes of the customer complaints and the average height of the multiple independent trees; and determining the root cause of the network complaint based on the abnormal indicators. The network complaint root cause location method provided in this application can locate abnormal indicators using an unsupervised method, without the need for pre-training an analysis model, and can quickly and efficiently identify the root cause of network complaints even for complaint types that have not appeared before.

Description

Technical Field

[0001] This application relates to the field of mobile communication technology, specifically to a method, apparatus, and application for locating the root cause of network complaints. Background Technology

[0002] With the development of 5G communication technology, the complexity of networks, service types, user types and data have increased dramatically. Network complaints involve many network performance indicators. If traditional manual location methods are used, the accuracy and efficiency of root cause location results are low.

[0003] Currently, there is a network fault root cause localization method in the existing technology. After determining that the network transmission anomaly is caused by the fault, it collects the operating data of each node in the network transmission and the network topology map. Based on the connection relationship between each node recorded in the topology map, it uses multiple analysis models in the model set to determine whether there are abnormal nodes among multiple nodes, and sequentially identifies the node that caused the network transmission anomaly and its root cause.

[0004] The above method requires the identification of abnormal nodes through multiple analysis models in the model set. In order to improve the accuracy and speed of network fault root cause localization, high requirements are placed on the number of analysis models in the model set and the accuracy of each analysis model. For complaint types that have not appeared before, new analysis models need to be retrained before the corresponding abnormal nodes and abnormal root causes can be identified and located. Summary of the Invention

[0005] This application provides a method, apparatus, and application for locating the root cause of online complaints, which solves the technical problem in the prior art that the model used for locating the root cause of online complaints needs to be trained in advance and is difficult to cover complaint types that have not appeared before.

[0006] In a first aspect, embodiments of this application provide a method for locating the root cause of online complaints, including:

[0007] Using the isolated forest algorithm, multiple independent trees are constructed based on a set of network performance indicators for customer complaints and a set of network performance indicators for non-customer complaints; the set of network performance indicators for customer complaints is the set of network performance indicators for customer complaint cells; the set of network performance indicators for non-customer complaint cells is the set of network performance indicators for non-customer complaint cells; the customer complaint cells are located within a preset range of the location where the network complaint was initiated.

[0008] Whether the customer complaint network performance index set is abnormal is determined based on the average height of the customer complaint network performance index set across the multiple independent trees;

[0009] When the set of customer complaint network performance indicators is abnormal, the abnormal indicator is determined based on the path height of the customer complaint leaf node and the average height of the multiple independent trees; the abnormal indicator is the segmentation benchmark of the abnormal leaf node; the proportion of the path height of the abnormal leaf node to the average height of the multiple independent trees is less than a preset value; the customer complaint leaf node is the leaf node in the multiple independent trees where the set of customer complaint network performance indicators is located.

[0010] The root cause of online complaints is determined based on the aforementioned abnormal indicators.

[0011] In one embodiment, determining whether the customer complaint network performance indicator set is abnormal based on the average height of the customer complaint network performance indicator set across the multiple independent trees includes:

[0012] The anomaly score of the customer complaint network performance index set is calculated based on the average height of the customer complaint network performance index set across the multiple independent trees;

[0013] When the abnormal score is greater than a preset abnormal score threshold, the set of customer complaint network performance indicators is determined to be abnormal.

[0014] In one embodiment, determining the anomaly index based on the path height of the complained leaf node and the average height of the multiple independent trees includes:

[0015] Calculate the average height E of the multiple independent trees. M ;

[0016] The path height is less than a×E M The complaint leaf node is used as the abnormal leaf node; where a∈(0,1);

[0017] The segmentation criterion of the abnormal leaf node is used as the abnormality index.

[0018] In one embodiment, the method of constructing multiple independent trees based on a set of customer complaint network performance indicators and a set of non-customer complaint network performance indicators using the isolated forest algorithm includes:

[0019] Construct a dataset containing N elements; where N-1 elements are the set of performance indicators for the non-complaint network, 1 element is the set of performance indicators for the complainant network, and N is an integer greater than 2;

[0020] Perform M element extractions on the dataset, and construct an independent tree for each extracted element to obtain the multiple independent trees; where M is an integer greater than 2.

[0021] In one embodiment, the dataset is subjected to M element extractions, and an independent tree is constructed for each extracted element, resulting in the multiple independent trees. The construction process of an independent tree is as follows:

[0022] Extract the building elements of the independent tree from the dataset at a preset sampling ratio;

[0023] The building element is split based on any one of the network performance metrics set as the splitting criterion, resulting in two branches;

[0024] For each branch's building element, the step of splitting the building element based on any index in the network performance index set is performed again until the tree splitting stopping condition is met, at which point the splitting of all building elements under all branches ends, and the independent tree is obtained.

[0025] Specifically, the splitting of the building elements under the current branch ends when the branch splitting stop condition is met; the branch splitting stop condition includes: there is exactly one building element under the current branch; the tree splitting stop condition includes: the splitting of building elements under all branches has ended or the height of the current tree is equal to or greater than the independent tree height threshold.

[0026] In one embodiment, before constructing multiple independent trees based on the customer complaint network performance index set and the non-customer complaint network performance index set using the isolated forest algorithm, the following steps are included:

[0027] The customer complaint cell is determined based on the location where the network complaint was initiated, and the set of customer complaint network performance indicators is obtained based on the customer complaint cell.

[0028] The distance between the customer complaint cell and the location where the online complaint was initiated is less than a preset distance threshold; the customer complaint cell is located within a preset azimuth angle range of the location where the online complaint was initiated.

[0029] In one embodiment, determining the root cause of the online complaint based on the abnormal indicators includes:

[0030] When the abnormal indicator meets one of the coverage abnormality conditions, the root cause of the network complaint is determined to include coverage abnormality. The coverage abnormality conditions include: there is no base station within a preset range of the location where the network complaint was initiated; the abnormal indicator includes an SA (Standalone) user coverage satisfaction indicator and the SA user coverage satisfaction indicator is less than the SA user coverage satisfaction threshold; and the abnormal indicator includes a measurement report MR (Mean Migration) coverage rate and the MR coverage rate of the resident cell is less than the MR coverage threshold. The resident cell is the cell ranked in the top K positions after sorting the cells according to the number of user equipment connections from largest to smallest; where K is a positive integer.

[0031] When the abnormal indicators meet one of the interference anomaly conditions, the root cause of the network complaint is determined to include interference anomaly; the interference anomaly conditions include: the abnormal indicators include overlapping coverage rate and the overlapping coverage rate is greater than the overlapping coverage rate threshold, and the abnormal indicators include interference noise floor indicator and the interference noise floor indicator is greater than the interference noise floor indicator threshold; the overlapping coverage rate threshold is the average overlapping coverage rate of the cells without customer complaints; the interference noise floor indicator threshold is the average interference noise floor indicator of the cells without customer complaints;

[0032] When the abnormal indicator meets the site failure conditions, the root cause of the network complaint is determined to include site failure; the site failure conditions include: there is a faulty site within a preset range of the location where the network complaint was initiated; the faulty site is a base station in the faulty site record within a preset time period before the time the network complaint was initiated.

[0033] When the abnormal indicators meet the capacity abnormality conditions, it is determined that the root cause of the network complaint includes capacity abnormality; the capacity abnormality conditions include: the abnormal indicators include a user count indicator and the user count indicator is greater than a user count threshold; the user count threshold is the average of the user count indicators of the cell with no customer complaints.

[0034] If the abnormal indicator does not meet any of the conditions of coverage abnormality, interference abnormality, site failure, and capacity abnormality, then the root cause of the network complaint is determined to be other abnormalities.

[0035] Secondly, embodiments of this application provide a device for locating the root cause of online complaints, comprising:

[0036] An independent tree construction module is used to: construct multiple independent trees based on a set of network performance indicators for customer complaints and a set of network performance indicators for non-customer complaints using the isolated forest algorithm; the set of network performance indicators for customer complaints is the set of network performance indicators for customer complaint cells; the set of network performance indicators for non-customer complaint cells is the set of network performance indicators for non-customer complaint cells; the customer complaint cells are located within a preset range of the location where the network complaint was initiated.

[0037] The anomaly determination module is used to: determine whether the customer complaint network performance index set is abnormal based on the average height of the customer complaint network performance index set in the multiple independent trees;

[0038] An abnormal indicator identification module is used to: determine the abnormal indicator based on the path height of the customer complaint network performance indicator set and the average height of the multiple independent trees when the customer complaint leaf node is abnormal; the abnormal indicator is the segmentation benchmark of the abnormal leaf node; the proportion of the path height of the abnormal leaf node to the average height of the multiple independent trees is less than a preset value; the customer complaint leaf node is the leaf node in the multiple independent trees where the customer complaint network performance indicator set is located.

[0039] The network complaint root cause matching module is used to: determine the root cause of network complaints based on the abnormal indicators.

[0040] Thirdly, embodiments of this application provide an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the program to implement the steps of the network complaint root cause localization method described in the first aspect.

[0041] Fourthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the steps of the network complaint root cause localization method described in the first aspect.

[0042] The network complaint root cause localization method provided in this application embodiment uses the Isolation Forest algorithm to construct a dataset by using the network performance index sets of customer complaint cells and non-customer complaint cells as independent trees. Since the Isolation Forest algorithm defines outliers as easily isolated outliers, i.e., sparsely distributed points far from high-density groups, this application embodiment uses the non-customer complaint network performance index set as the high-density group. Based on the average height of the customer complaint network performance index set across multiple independent trees, the distance between the customer complaint network performance index set and the high-density group can be determined, thereby judging whether the customer complaint network performance index set is abnormal. After determining that the customer complaint network performance index set is abnormal, further analysis is performed based on the path height of the customer complaint leaf nodes. Anomalies were identified by analyzing the degree and average height of the independent trees. The proportion of the path height of the abnormal leaf node to the average height of multiple independent trees was less than a preset value, indicating a significant difference in the set of network performance indicators between the complained and non-complained cells when the abnormal leaf node was segmented. Therefore, the segmentation benchmark of the abnormal leaf node can be regarded as an anomaly indicator. The above process uses an unsupervised method to locate anomalies, thus eliminating the need for pre-training of the analysis model. Once the anomaly of a particular indicator is determined, the influencing factors related to that indicator can be identified as the root cause of network complaints. Since the correspondence between indicators and influencing factors is readily available, the root cause of network complaints can also be quickly and efficiently identified for complaint types that have not appeared before. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a flowchart illustrating the method for locating the root cause of online complaints provided in an embodiment of this application.

[0045] Figure 2This is a flowchart illustrating the independent tree construction process provided in an embodiment of this application;

[0046] Figure 3 This is a flowchart illustrating the method for determining whether a set of customer complaint network performance indicators is abnormal, as provided in an embodiment of this application.

[0047] Figure 4 This is a schematic diagram of the structure of the network complaint root cause localization device provided in the embodiments of this application;

[0048] Figure 5 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0050] Figure 1 This is a flowchart illustrating the method for locating the root cause of online complaints provided in an embodiment of this application. (Refer to...) Figure 1 This application provides a method for locating the root cause of online complaints, which may include:

[0051] S11. Using the isolated forest algorithm, construct multiple independent trees based on the customer complaint network performance index set and the non-customer complaint network performance index set;

[0052] S12. Determine whether the customer complaint network performance index set is abnormal based on the average height of the customer complaint network performance index set in multiple independent trees;

[0053] S13. When the set of customer complaint network performance indicators is abnormal, the abnormal indicators are determined based on the path height of the customer complaint leaf node and the average height of multiple independent trees.

[0054] S14. Determine the root cause of online complaints based on abnormal indicators.

[0055] In step S11, the set of network performance indicators for customer complaints is the set of network performance indicators for the customer complaint cells; the set of network performance indicators for non-customer complaint cells is the set of network performance indicators for the non-customer complaint cells.

[0056] The number of communities with customer complaints can be determined based on the location where the online complaint was initiated. The community with customer complaints is located within a preset range of the location where the online complaint was initiated. The preset range can include two dimensions: distance and azimuth. Correspondingly, a community that is not located within the preset range of any online complaint initiation location is a community with no customer complaints.

[0057] Furthermore, both the customer complaint network performance index set and the non-customer complaint network performance index set are multiple network performance index data based on a preset number of days prior to the time the network complaint was initiated, including but not limited to: 5G call success rate, wireless drop rate, SA handover success rate, standalone network user coverage satisfaction, MR coverage rate, overlapping coverage rate, interference noise floor index, and user number index.

[0058] In step S12, since outliers that are easily isolated in the isolated forest algorithm are often located in the smaller height of the independent trees, it is possible to determine whether the customer complaint network performance index set is easily isolated in the multiple independent trees based on the average height of the customer complaint network performance index set in multiple independent trees, and thus determine whether the customer complaint network performance index set is abnormal.

[0059] In step S13, the path height of the complained leaf node in each independent tree can be compared with the average height of multiple independent trees to determine whether the complained leaf node is an abnormal leaf node. The abnormality index is the segmentation benchmark for abnormal leaf nodes.

[0060] Among them, abnormal leaf nodes are customer complaint leaf nodes whose path height is less than the average height of the multiple independent trees; customer complaint leaf nodes are leaf nodes in the multiple independent trees where the customer complaint network performance index set is located.

[0061] In step S14, it is believed that the formation of online complaints is affected by abnormal indicators. Therefore, the corresponding influencing factors can be found based on the abnormal indicators, and the cause of the abnormal indicators can be determined, that is, the root cause of online complaints can be determined.

[0062] It should be noted that there can be one or more root causes for online complaints, and the embodiments of this application do not limit the number of root causes for online complaints.

[0063] The network complaint root cause localization method provided in this application embodiment uses the Isolation Forest algorithm to construct a dataset by using the network performance index sets of customer complaint cells and non-customer complaint cells as independent trees. Since the Isolation Forest algorithm defines outliers as easily isolated outliers, i.e., sparsely distributed points far from high-density groups, this application embodiment uses the non-customer complaint network performance index set as the high-density group. Based on the average height of the customer complaint network performance index set across multiple independent trees, the distance between the customer complaint network performance index set and the high-density group can be determined, thereby judging whether the customer complaint network performance index set is abnormal. After determining that the customer complaint network performance index set is abnormal, further analysis is performed based on the path height of the customer complaint leaf nodes. Anomalies were identified by analyzing the degree and average height of the independent trees. The proportion of the path height of the abnormal leaf node to the average height of multiple independent trees was less than a preset value, indicating a significant difference in the set of network performance indicators between the complained and non-complained cells when the abnormal leaf node was segmented. Therefore, the segmentation benchmark of the abnormal leaf node can be regarded as an anomaly indicator. The above process uses an unsupervised method to locate anomalies, thus eliminating the need for pre-training of the analysis model. Once the anomaly of a particular indicator is determined, the influencing factors related to that indicator can be identified as the root cause of network complaints. Since the correspondence between indicators and influencing factors is readily available, the root cause of network complaints can also be quickly and efficiently identified for complaint types that have not appeared before.

[0064] In one embodiment, before constructing an independent tree, the customer complaint cell can be determined based on the location where the network complaint was initiated, and a set of customer complaint network performance indicators can be obtained based on the customer complaint cell.

[0065] Wherein, the distance between the customer complaint cell and the location where the online complaint was initiated is less than a preset distance threshold; and the customer complaint cell is located within a preset azimuth angle range of the location where the online complaint was initiated.

[0066] In the process of identifying the aforementioned residential communities with customer complaints, the distance between the location where the online complaint was initiated and the community was calculated using the Haversine formula:

[0067]

[0068] Where d represents the distance between the location where the online complaint was initiated and the community; r represents the Earth's radius; λ1 and These represent the longitude and latitude of the location where the online complaint was initiated; λ2 and These represent the longitude and latitude of the community, respectively.

[0069] In the process of identifying the aforementioned residential communities with customer complaints, the azimuth angle between the location where the online complaint was initiated and the community can be calculated using the following formula:

[0070]

[0071] Where θ represents the azimuth angle between the location where the online complaint was initiated and the community.

[0072] When d is less than the preset distance threshold and θ is within the preset azimuth angle range, it indicates that the current cell is the cell corresponding to the current network complaint, i.e., the customer complaint cell; correspondingly, cells that are not within the preset range of any network complaint initiation location are customer complaint cells, which means that no abnormalities were found in the network performance index set of customer complaint cells.

[0073] It should be noted that when identifying customer complaint communities, there may be situations where multiple communities are located within the preset range of the location where the online complaint was initiated. In this case, the communities can be sorted in ascending order of their distance from the location where the online complaint was initiated, and the community ranked in the top P positions is taken as the customer complaint community, where P is a positive integer less than or equal to 10.

[0074] It is understood that the embodiments of this application do not have a unique limitation on the preset distance threshold and the preset azimuth range. In actual application, the specific values ​​of the preset distance threshold and the preset azimuth range can be set according to actual needs.

[0075] The customer complaint cell identification method provided in this application uses the distance and azimuth information between the location where the network complaint was initiated and the cell to identify the customer complaint cell within a preset range of the location where the network complaint was initiated. By using the two dimensions of distance and azimuth as the judgment criteria, the accuracy of customer complaint cell location can be improved, thereby obtaining an accurate set of customer complaint network performance indicators to ensure the accuracy of root cause location of network complaints.

[0076] In one embodiment, see Figure 2 The process of constructing multiple independent trees based on the customer complaint network performance index set and the non-customer complaint network performance index set using the Isolation Forest algorithm is as follows:

[0077] S21. Construct a dataset containing N elements;

[0078] S22. Perform M element extractions on the dataset and construct an independent tree for each extracted element to obtain multiple independent trees.

[0079] In step S21, the N elements can be composed of N-1 sets of network performance indicators without customer complaints and 1 set of network performance indicators with customer complaints.

[0080] Where N is an integer greater than 2; for example, N can be set to 1000.

[0081] In step S22, M is an integer greater than 2; for example, M can be set to 100.

[0082] In one embodiment, the construction process of the independent tree in step S22 is illustrated using an independent tree as an example. The specific steps are as follows:

[0083] Extract the building elements of the independent tree from the dataset at a preset sampling ratio;

[0084] The building element is split based on any one of the network performance metrics set as the splitting criterion, resulting in two branches;

[0085] For each branch's building element, the step of splitting the building element based on any index in the network performance index set is performed again until the tree splitting stopping condition is met, at which point the splitting of all building elements under all branches ends, and the independent tree is obtained.

[0086] Assuming that in the above process, the number of elements in the dataset is 1000, the number of independent trees constructed is 100, and the preset sampling ratio can be set to 50%, that is, 500 construction elements are extracted each time to construct an independent tree, that is, each independent tree consists of 500 construction elements.

[0087] It should be noted that the above-mentioned preset sampling ratio of 50% is an example in the embodiments of this application and does not constitute the only limitation on the embodiments of this application. The specific value of the preset sampling ratio can be set according to the actual situation.

[0088] The process of segmenting the building elements based on any one of the network performance metrics as the segmentation criterion is as follows:

[0089] In the set of network performance metrics, any metric is selected as the splitting benchmark for the current node. Any value from the minimum to the maximum value of the metric in the building element is used as the benchmark value. The value of the metric for each building element is compared with the benchmark value. If it is less than the benchmark value, it is considered as one branch, and if it is greater than or equal to the benchmark value, it is considered as another branch. Thus, the building element is divided into two branches.

[0090] It should be noted that during the process of further splitting the build elements under each branch, the branch will trigger a branch splitting stop condition, thus prematurely ending the splitting action of the current branch. Specifically, the splitting of the build elements under the current branch ends when the current branch meets the branch splitting stop condition. The branch splitting stop condition includes: the current branch has exactly one build element.

[0091] When there is only one building element in a branch, the remaining building elements no longer meet the minimum number of elements required to perform the split operation, so the split operation of the current branch is stopped.

[0092] In this embodiment of the application, the determination condition for the completion of independent tree construction is the tree segmentation stopping condition, which includes: the segmentation of all construction elements under all branches has been completed or the height of the current tree is equal to or greater than the independent tree height threshold.

[0093] It should be noted that the setting of the independent tree height threshold in this embodiment can be set according to the actual situation, and is not limited to a single value here.

[0094] Since this application embodiment uses the isolated forest algorithm to identify outliers that are easily isolated, and then finds anomaly indicators, outliers in the isolated tree are often located in leaf nodes with small path heights. Therefore, leaf nodes with small path heights are more likely to be outliers, while normal points with long path heights are not key points in the isolated forest algorithm. Therefore, the height of the isolated tree is limited. When the height of the current tree is equal to or greater than the height threshold of the isolated tree, it can be regarded as normal points for subsequent segmentation. Ending the segmentation at this time can save a lot of unnecessary segmentation work.

[0095] The independent tree construction method provided in this application constructs multiple independent trees through sampling and extraction. Sampling and extraction reduces the data processing volume of each independent tree, improves the construction speed of a single independent tree, and obtains a sufficient data set for the judgment and determination of abnormal customer complaint network performance indicators by extracting elements M times and constructing independent trees M times. This avoids the deviation in the results of the judgment and determination of abnormal customer complaint network performance indicators due to errors introduced by the small data set.

[0096] In one embodiment, see Figure 3 Step S12 may include:

[0097] S31. Calculate the anomaly score of the customer complaint network performance index set based on the average height of the customer complaint network performance index set in multiple independent trees;

[0098] S32. When the abnormal score is greater than the preset abnormal score threshold, the set of customer complaint network performance indicators is determined to be abnormal.

[0099] In step S31, the anomaly score s(X) of the customer complaint network performance index set X can be calculated according to the following formula:

[0100]

[0101] Where E(h(X)) represents the average height of the customer complaint network performance index set across multiple independent trees, and N represents the number of elements in the dataset.

[0102] In the above calculation formula, c(N) is calculated according to the following formula:

[0103]

[0104] Where H represents the harmonic function.

[0105] When the abnormal score is greater than the preset abnormal score threshold, it indicates that the set of customer complaint network performance indicators is likely to be isolated and is considered abnormal; when the abnormal score is less than or equal to the preset abnormal score threshold, it indicates that the set of customer complaint network performance indicators is likely to be isolated and is considered normal, and there is no need to locate the root cause of network complaints.

[0106] In this embodiment of the application, the preset abnormal score threshold can be set to 0.8. It should be noted that the above-mentioned preset abnormal score threshold is only an example value. In actual applications, the preset abnormal score threshold can be adjusted according to the actual situation.

[0107] In one embodiment, the method for determining abnormal indicators is as follows:

[0108] Calculate the average height E of multiple independent trees M ;

[0109] The path height is less than a×E M The customer complaint leaf node is used as the abnormal leaf node; where a∈(0,1); for example, a can take the value 0.2;

[0110] The segmentation criterion for abnormal leaf nodes is used as the anomaly indicator.

[0111] In this embodiment, for the set of customer complaint network performance indicators whose anomaly scores are greater than a preset anomaly score threshold, the system filters out those belonging to the customer complaint network performance indicator set and whose path length is less than a×E from all independent trees. M The leaf nodes with customer complaints are considered as abnormal leaf nodes. At this time, the leaf nodes with customer complaints are located in the nodes that are preferentially isolated in the independent tree. This indicates that the set of network performance indicators with customer complaints can be easily distinguished from the set of network performance indicators without customer complaints based on the segmentation benchmark corresponding to the abnormal leaf nodes. This indicates that when the abnormal leaf nodes are segmented, there is a significant difference in the set of network performance indicators between the cells with customer complaints and the cells without customer complaints. Therefore, the segmentation benchmark of the abnormal leaf nodes can be regarded as abnormal indicators.

[0112] The method for determining abnormal indicators provided in this application embodiment can identify abnormal leaf nodes in customer complaint leaf nodes based solely on the path height of the leaf node and the average height of multiple independent trees. Then, it determines abnormal indicators based on the segmentation benchmark of the abnormal leaf node. Compared with the network complaint root cause localization method that uses analysis models in the prior art, it simplifies the model training steps and reduces the requirements for the accuracy and number of analysis models, making the network complaint root cause localization process more efficient and convenient.

[0113] In one embodiment, the root causes of network complaints can be of various types, including but not limited to: coverage anomalies, interference anomalies, site failures, and capacity anomalies.

[0114] The root cause of online complaints may be one or more. Therefore, this application proposes a method for determining the corresponding root cause of online complaints based on the conditions met by abnormal indicators. The process of determining the root cause of online complaints based on abnormal indicators is described below:

[0115] When the abnormal indicator meets one of the coverage abnormality conditions, the root cause of the network complaint is determined to include coverage abnormality. The coverage abnormality conditions include: there is no base station within a preset range of the location where the network complaint was initiated; the abnormal indicator includes an SA (Standalone) user coverage satisfaction indicator and the SA user coverage satisfaction indicator is less than the SA user coverage satisfaction threshold; and the abnormal indicator includes a measurement report MR (Mean Migration) coverage rate and the MR coverage rate of the resident cell is less than the MR coverage threshold. The resident cell is the cell ranked in the top K positions after sorting the cells according to the number of user equipment connections from largest to smallest; where K is a positive integer.

[0116] When the abnormal indicators meet one of the interference anomaly conditions, the root cause of the network complaint is determined to include interference anomaly; the interference anomaly conditions include: the abnormal indicators include overlapping coverage rate and the overlapping coverage rate is greater than the overlapping coverage rate threshold, and the abnormal indicators include interference noise floor indicator and the interference noise floor indicator is greater than the interference noise floor indicator threshold; the overlapping coverage rate threshold is the average overlapping coverage rate of the cells without customer complaints; the interference noise floor indicator threshold is the average interference noise floor indicator of the cells without customer complaints;

[0117] When the abnormal indicator meets the site failure conditions, the root cause of the network complaint is determined to include site failure; the site failure conditions include: there is a faulty site within a preset range of the location where the network complaint was initiated; the faulty site is a base station in the faulty site record within a preset time period before the time the network complaint was initiated.

[0118] When the abnormal indicators meet the capacity abnormality conditions, it is determined that the root cause of the network complaint includes capacity abnormality; the capacity abnormality conditions include: the abnormal indicators include a user count indicator and the user count indicator is greater than a user count threshold; the user count threshold is the average of the user count indicators of the cell with no customer complaints.

[0119] If the abnormal indicator does not meet any of the conditions of coverage abnormality, interference abnormality, site failure, and capacity abnormality, then the root cause of the network complaint is determined to be other abnormalities.

[0120] It should be noted that the network complaint in this application embodiment can simultaneously meet multiple conditions among coverage anomaly, interference anomaly, site failure, and capacity anomaly. Whenever one of the conditions is met, it indicates that the root cause of the network complaint includes the cause corresponding to that condition. If the network complaint does not meet any of the conditions among coverage anomaly, interference anomaly, site failure, and capacity anomaly, but anomalies are identified in the customer complaint network performance indicator set, it indicates that the network complaint has a corresponding cause, but does not belong to any of the coverage anomaly, interference anomaly, site failure, and capacity anomaly. In this case, the root cause of the network complaint is determined to be other anomalies.

[0121] In the above process, the overlap coverage threshold, interference noise floor index threshold, and user number threshold are calculated through a set of network performance indicators without customer complaints. Compared with fixed preset values, these are more accurate and reasonable, and can improve the accuracy and reliability of determining the root cause of network complaints.

[0122] Furthermore, the MR coverage threshold and the SA user coverage satisfaction threshold can also be calculated based on a set of network performance indicators without customer complaints.

[0123] Furthermore, in this embodiment of the application, various anomaly types can be correlated with the conditions met by indicators to form a root cause condition mapping table. When executing step S14, the conditions met by the anomaly indicators are searched in the root cause condition mapping table to find the corresponding anomaly type as the root cause of the network complaint.

[0124] The correspondence between various anomaly types and the conditions for meeting the indicators can be determined based on the correspondence between the indicators and the influencing factors, or the root cause condition mapping table can be updated by manually adding them.

[0125] The network complaint root cause localization method provided in this application retrieves the conditions that the abnormal indicators meet, and thus finds the network complaint root cause corresponding to the conditions met. That is, when it is determined which indicator is abnormal, the influencing factors related to that indicator can be identified as the network complaint root cause. Since the correspondence between indicators and influencing factors is easily obtainable information, the network complaint root cause can also be quickly and efficiently identified for complaint types that have not appeared before.

[0126] The following describes the network complaint root cause localization device provided in the embodiments of this application. The network complaint root cause localization device described below can be referred to in correspondence with the network complaint root cause localization method described above.

[0127] See Figure 4 The network complaint root cause localization device provided in this application embodiment includes:

[0128] The independent tree construction module 410 is used to: construct multiple independent trees based on the customer complaint network performance index set and the non-customer complaint network performance index set using the isolated forest algorithm; the customer complaint network performance index set is the network performance index set of the customer complaint cell; the non-customer complaint network performance index set is the network performance index set of the non-customer complaint cell; the customer complaint cell is located within a preset range of the location where the network complaint was initiated.

[0129] The anomaly determination module 420 is used to: determine whether the customer complaint network performance index set is abnormal based on the average height of the customer complaint network performance index set in the multiple independent trees;

[0130] The abnormal indicator identification module 430 is used to: determine the abnormal indicator based on the path height of the customer complaint network performance indicator set and the average height of the multiple independent trees when the customer complaint leaf node is abnormal; the abnormal indicator is the segmentation benchmark of the abnormal leaf node; the proportion of the path height of the abnormal leaf node to the average height of the multiple independent trees is less than a preset value; the customer complaint leaf node is the leaf node in the multiple independent trees where the customer complaint network performance indicator set is located.

[0131] The network complaint root cause matching module 440 is used to: determine the root cause of the network complaint based on the abnormal indicators.

[0132] In the network complaint root cause localization device provided in this application embodiment, the independent tree construction module constructs multiple independent trees based on the network performance index sets of the complained cell and the non-complained cell. Through the anomaly judgment module, the distance between the complained network performance index set and the high-density group can be determined by combining the average height of the complained network performance index set in the multiple independent trees, thereby determining whether the complained network performance index set is abnormal. After determining that the complained network performance index set is abnormal, the anomaly indicator identification module further finds the abnormal indicators based on the path height of the complained leaf node and the average height of the independent tree. The proportion of the path height of the abnormal leaf node to the average height of the multiple independent trees is less than a preset value, indicating that there is a significant difference in the network performance index sets between the complained cell and the non-complained cell when the abnormal leaf node is segmented. Therefore, the segmentation benchmark of the abnormal leaf node can be regarded as an abnormal indicator. The network complaint root cause matching module determines the network complaint root cause based on the abnormal indicator. In this process, an unsupervised method is used to locate the abnormal indicator root cause without pre-training the analysis model. Moreover, for complaint types that have not appeared before, the network complaint root cause can also be quickly and efficiently identified based on the abnormal indicator and its related influencing factors.

[0133] In one embodiment, the independent tree building module 410 is specifically used for:

[0134] Construct a dataset containing N elements; where N-1 elements are the set of performance indicators for the non-complaint network, 1 element is the set of performance indicators for the complainant network, and N is an integer greater than 2;

[0135] Perform M element extractions on the dataset, and construct an independent tree for each extracted element to obtain the multiple independent trees; where M is an integer greater than 2.

[0136] In one embodiment, the process by which the independent tree construction module 410 constructs an independent tree is as follows:

[0137] Extract the building elements of the independent tree from the dataset at a preset sampling ratio;

[0138] The building element is split based on any one of the network performance metrics set as the splitting criterion, resulting in two branches;

[0139] For each branch's building element, the step of splitting the building element based on any index in the network performance index set is performed again until the tree splitting stopping condition is met, at which point the splitting of all building elements under all branches ends, and the independent tree is obtained.

[0140] Specifically, the splitting of the building elements under the current branch ends when the branch splitting stop condition is met; the branch splitting stop condition includes: there is exactly one building element under the current branch; the tree splitting stop condition includes: the splitting of building elements under all branches has ended or the height of the current tree is equal to or greater than the independent tree height threshold.

[0141] In one embodiment, the anomaly detection module 420 is specifically used for:

[0142] The anomaly score of the customer complaint network performance index set is calculated based on the average height of the customer complaint network performance index set across the multiple independent trees;

[0143] When the abnormal score is greater than a preset abnormal score threshold, the set of customer complaint network performance indicators is determined to be abnormal.

[0144] In one embodiment, the abnormal indicator identification module 430 is specifically used for:

[0145] Calculate the average height E of the multiple independent trees. M ;

[0146] The path height is less than a×E M The complaint leaf node is used as the abnormal leaf node; where a∈(0,1);

[0147] The segmentation criterion of the abnormal leaf node is used as the abnormality index.

[0148] In one embodiment, the online complaint root cause matching module 440 is specifically used for:

[0149] When the abnormal indicator meets one of the coverage abnormality conditions, the root cause of the network complaint is determined to include coverage abnormality. The coverage abnormality conditions include: there is no base station within a preset range of the location where the network complaint was initiated; the abnormal indicator includes an SA (Standalone) user coverage satisfaction indicator and the SA user coverage satisfaction indicator is less than the SA user coverage satisfaction threshold; and the abnormal indicator includes a measurement report MR (Mean Migration) coverage rate and the MR coverage rate of the resident cell is less than the MR coverage threshold. The resident cell is the cell ranked in the top K positions after sorting the cells according to the number of user equipment connections from largest to smallest; where K is a positive integer.

[0150] When the abnormal indicators meet one of the interference anomaly conditions, the root cause of the network complaint is determined to include interference anomaly; the interference anomaly conditions include: the abnormal indicators include overlapping coverage rate and the overlapping coverage rate is greater than the overlapping coverage rate threshold, and the abnormal indicators include interference noise floor indicator and the interference noise floor indicator is greater than the interference noise floor indicator threshold; the overlapping coverage rate threshold is the average overlapping coverage rate of the cells without customer complaints; the interference noise floor indicator threshold is the average interference noise floor indicator of the cells without customer complaints;

[0151] When the abnormal indicator meets the site failure conditions, the root cause of the network complaint is determined to include site failure; the site failure conditions include: there is a faulty site within a preset range of the location where the network complaint was initiated; the faulty site is a base station in the faulty site record within a preset time period before the time the network complaint was initiated.

[0152] When the abnormal indicators meet the capacity abnormality conditions, it is determined that the root cause of the network complaint includes capacity abnormality; the capacity abnormality conditions include: the abnormal indicators include a user count indicator and the user count indicator is greater than a user count threshold; the user count threshold is the average of the user count indicators of the cell with no customer complaints.

[0153] If the abnormal indicator does not meet any of the conditions of coverage abnormality, interference abnormality, site failure, and capacity abnormality, then the root cause of the network complaint is determined to be other abnormalities.

[0154] In one embodiment, the network complaint root cause localization device provided in this application further includes a customer complaint cell identification module (not shown in the figure), used for:

[0155] The customer complaint cell is determined based on the location where the network complaint was initiated, and the set of customer complaint network performance indicators is obtained based on the customer complaint cell; wherein, the distance between the customer complaint cell and the location where the network complaint was initiated is less than a preset distance threshold; and the customer complaint cell is located within a preset azimuth angle range of the location where the network complaint was initiated.

[0156] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other via the communication bus 540. The processor 510 can call a computer program in the memory 530 to execute the steps of the network complaint root cause localization method, such as including:

[0157] Using the isolated forest algorithm, multiple independent trees are constructed based on a set of network performance indicators for customer complaints and a set of network performance indicators for non-customer complaints; the set of network performance indicators for customer complaints is the set of network performance indicators for customer complaint cells; the set of network performance indicators for non-customer complaint cells is the set of network performance indicators for non-customer complaint cells; the customer complaint cells are located within a preset range of the location where the network complaint was initiated.

[0158] Whether the customer complaint network performance index set is abnormal is determined based on the average height of the customer complaint network performance index set across the multiple independent trees;

[0159] When the set of customer complaint network performance indicators is abnormal, the abnormal indicator is determined based on the path height of the customer complaint leaf node and the average height of the multiple independent trees; the abnormal indicator is the segmentation benchmark of the abnormal leaf node; the proportion of the path height of the abnormal leaf node to the average height of the multiple independent trees is less than a preset value; the customer complaint leaf node is the leaf node in the multiple independent trees where the set of customer complaint network performance indicators is located.

[0160] The root cause of online complaints is determined based on the aforementioned abnormal indicators.

[0161] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0162] On the other hand, embodiments of this application also provide a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can perform the steps of the network complaint root cause localization method provided in the above embodiments, such as including:

[0163] Using the isolated forest algorithm, multiple independent trees are constructed based on a set of network performance indicators for customer complaints and a set of network performance indicators for non-customer complaints; the set of network performance indicators for customer complaints is the set of network performance indicators for customer complaint cells; the set of network performance indicators for non-customer complaint cells is the set of network performance indicators for non-customer complaint cells; the customer complaint cells are located within a preset range of the location where the network complaint was initiated.

[0164] Whether the customer complaint network performance index set is abnormal is determined based on the average height of the customer complaint network performance index set across the multiple independent trees;

[0165] When the set of customer complaint network performance indicators is abnormal, the abnormal indicator is determined based on the path height of the customer complaint leaf node and the average height of the multiple independent trees; the abnormal indicator is the segmentation benchmark of the abnormal leaf node; the proportion of the path height of the abnormal leaf node to the average height of the multiple independent trees is less than a preset value; the customer complaint leaf node is the leaf node in the multiple independent trees where the set of customer complaint network performance indicators is located.

[0166] The root cause of online complaints is determined based on the aforementioned abnormal indicators.

[0167] On the other hand, embodiments of this application also provide a processor-readable storage medium storing a computer program for causing a processor to perform the steps of the methods provided in the above embodiments, such as including:

[0168] Using the isolated forest algorithm, multiple independent trees are constructed based on a set of network performance indicators for customer complaints and a set of network performance indicators for non-customer complaints; the set of network performance indicators for customer complaints is the set of network performance indicators for customer complaint cells; the set of network performance indicators for non-customer complaint cells is the set of network performance indicators for non-customer complaint cells; the customer complaint cells are located within a preset range of the location where the network complaint was initiated.

[0169] Whether the customer complaint network performance index set is abnormal is determined based on the average height of the customer complaint network performance index set across the multiple independent trees;

[0170] When the set of customer complaint network performance indicators is abnormal, the abnormal indicator is determined based on the path height of the customer complaint leaf node and the average height of the multiple independent trees; the abnormal indicator is the segmentation benchmark of the abnormal leaf node; the proportion of the path height of the abnormal leaf node to the average height of the multiple independent trees is less than a preset value; the customer complaint leaf node is the leaf node in the multiple independent trees where the set of customer complaint network performance indicators is located.

[0171] The root cause of online complaints is determined based on the aforementioned abnormal indicators.

[0172] The processor-readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to magnetic memory (e.g., floppy disk, hard disk, magnetic tape, magneto-optical disk (MO)), optical memory (e.g., CD, DVD, BD, HVD), and semiconductor memory (e.g., ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD)).

[0173] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0174] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, 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 can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0175] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for locating the root cause of online complaints, characterized in that, include: Using the isolated forest algorithm, multiple independent trees are constructed based on a set of network performance indicators for customer complaints and a set of network performance indicators for non-customer complaints; the set of network performance indicators for customer complaints is the set of network performance indicators for customer complaint cells; the set of network performance indicators for non-customer complaint cells is the set of network performance indicators for non-customer complaint cells; the customer complaint cells are located within a preset range of the location where the network complaint was initiated. Whether the customer complaint network performance index set is abnormal is determined based on the average height of the customer complaint network performance index set across the multiple independent trees; When the set of customer complaint network performance indicators is abnormal, the abnormal indicator is determined based on the path height of the customer complaint leaf node and the average height of the multiple independent trees; the abnormal indicator is the segmentation benchmark for the abnormal leaf node; the proportion of the path height of the abnormal leaf node to the average height of the multiple independent trees is less than a preset value; the customer complaint leaf node is the leaf node in the multiple independent trees where the set of customer complaint network performance indicators is located. The root cause of online complaints is determined based on the aforementioned abnormal indicators.

2. The method for locating the root cause of online complaints according to claim 1, characterized in that, The step of determining whether the customer complaint network performance indicator set is abnormal based on the average height of the customer complaint network performance indicator set across the multiple independent trees includes: The anomaly score of the customer complaint network performance index set is calculated based on the average height of the customer complaint network performance index set across the multiple independent trees; When the abnormal score is greater than a preset abnormal score threshold, the set of customer complaint network performance indicators is determined to be abnormal.

3. The method for locating the root cause of online complaints according to claim 1, characterized in that, The step of determining the abnormal indicators based on the path height of the leaf node in the customer complaint and the average height of the multiple independent trees includes: calculating an average height E of the plurality of independent trees M ; The path height is less than a×E M The complaint leaf node is used as the abnormal leaf node; where a∈(0,1); The segmentation criterion of the abnormal leaf node is used as the abnormality index.

4. The method for locating the root cause of online complaints according to claim 1, characterized in that, The method utilizes the isolated forest algorithm to construct multiple independent trees based on a set of network performance indicators for customers and a set of network performance indicators for those without customer complaints, including: Construct a dataset containing N elements; where N-1 elements are the set of performance indicators for the non-complaint network, 1 element is the set of performance indicators for the complainant network, and N is an integer greater than 2; Perform M element extractions on the dataset, and construct an independent tree for each extracted element to obtain the multiple independent trees; where M is an integer greater than 2.

5. The method for locating the root cause of online complaints according to claim 4, characterized in that, The dataset is subjected to M element extractions, and an independent tree is constructed for each extracted element, resulting in the multiple independent trees. The construction process of an independent tree is as follows: Extract the building elements of the independent tree from the dataset at a preset sampling ratio; The building element is split based on any one of the network performance metrics set as the splitting criterion, resulting in two branches; For each branch's building element, the step of splitting the building element based on any index in the network performance index set is performed again until the tree splitting stopping condition is met, at which point the splitting of all building elements under all branches ends, and the independent tree is obtained. Specifically, the splitting of the building elements under the current branch ends when the branch splitting stop condition is met; the branch splitting stop condition includes: there is exactly one building element under the current branch; the tree splitting stop condition includes: the splitting of building elements under all branches has ended or the height of the current tree is equal to or greater than the independent tree height threshold.

6. The method for locating the root cause of online complaints according to claim 1, characterized in that, Before constructing multiple independent trees based on the customer complaint network performance index set and the non-customer complaint network performance index set using the isolated forest algorithm, the following steps are included: The customer complaint cell is determined based on the location where the network complaint was initiated, and the set of customer complaint network performance indicators is obtained based on the customer complaint cell. The distance between the customer complaint cell and the location where the online complaint was initiated is less than a preset distance threshold; the customer complaint cell is located within a preset azimuth angle range of the location where the online complaint was initiated.

7. The method for locating the root cause of online complaints according to claim 1, characterized in that, The step of determining the root cause of online complaints based on the abnormal indicators includes: When the abnormal indicator meets one of the coverage abnormality conditions, the root cause of the network complaint is determined to include coverage abnormality. The coverage abnormality conditions include: there is no base station within a preset range of the location where the network complaint was initiated; the abnormal indicator includes an SA (Standalone) user coverage satisfaction indicator and the SA user coverage satisfaction indicator is less than the SA user coverage satisfaction threshold; and the abnormal indicator includes a measurement report MR (Mean Migration) coverage rate and the MR coverage rate of the resident cell is less than the MR coverage threshold. The resident cell is the cell ranked in the top K positions after sorting the cells according to the number of user equipment connections from largest to smallest; where K is a positive integer. When the abnormal indicators meet one of the interference anomaly conditions, the root cause of the network complaint is determined to include interference anomaly; the interference anomaly conditions include: the abnormal indicators include overlapping coverage rate and the overlapping coverage rate is greater than the overlapping coverage rate threshold, and the abnormal indicators include interference noise floor indicator and the interference noise floor indicator is greater than the interference noise floor indicator threshold; the overlapping coverage rate threshold is the average overlapping coverage rate of the cells without customer complaints; the interference noise floor indicator threshold is the average interference noise floor indicator of the cells without customer complaints; When the abnormal indicator meets the site failure conditions, the root cause of the network complaint is determined to include site failure; the site failure conditions include: there is a faulty site within a preset range of the location where the network complaint was initiated; the faulty site is a base station in the faulty site record within a preset time period before the time the network complaint was initiated. When the abnormal indicators meet the capacity abnormality conditions, it is determined that the root cause of the network complaint includes capacity abnormality; the capacity abnormality conditions include: the abnormal indicators include a user count indicator and the user count indicator is greater than a user count threshold; the user count threshold is the average of the user count indicators of the cell with no customer complaints. If the abnormal indicator does not meet any of the conditions of coverage abnormality, interference abnormality, site failure, and capacity abnormality, then the root cause of the network complaint is determined to be other abnormalities.

8. A device for locating the root cause of online complaints, characterized in that, include: An independent tree construction module is used to: construct multiple independent trees based on a set of network performance indicators for customer complaints and a set of network performance indicators for non-customer complaints using the isolated forest algorithm; the set of network performance indicators for customer complaints is the set of network performance indicators for customer complaint cells; the set of network performance indicators for non-customer complaint cells is the set of network performance indicators for non-customer complaint cells; the customer complaint cells are located within a preset range of the location where the network complaint was initiated. The anomaly determination module is used to: determine whether the customer complaint network performance index set is abnormal based on the average height of the customer complaint network performance index set in the multiple independent trees; An abnormal indicator identification module is used to: determine the abnormal indicator based on the path height of the customer complaint network performance indicator set and the average height of the multiple independent trees when the customer complaint leaf node is abnormal; the abnormal indicator is the segmentation benchmark of the abnormal leaf node; the proportion of the path height of the abnormal leaf node to the average height of the multiple independent trees is less than a preset value; the customer complaint leaf node is the leaf node in the multiple independent trees where the customer complaint network performance indicator set is located. The network complaint root cause matching module is used to: determine the root cause of network complaints based on the abnormal indicators.

9. An electronic device comprising a processor and a memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the network complaint root cause localization method according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the network complaint root cause localization method according to any one of claims 1 to 7.

Images