Method and system for positioning abnormal root cause of user communication by fusing multi-dimensional data

Abstract

The invention provides a method and system for positioning a user communication exception root cause by fusing multi-dimensional data, and belongs to the technical field of mobile communication and big data. The method comprises the following steps of: searching a corresponding user call record by taking a user ID (Identity) of an abnormal call as a trigger; determining a user call record key data item sub-root cause and a corresponding optimization scheme; determining network elements related to the abnormal call, setting feature data items of the network elements, and obtaining sub-root causes of the feature data items of the corresponding network elements and corresponding optimization schemes; setting the weight of each sub-root cause, and merging the sub-root causes according to a sub-root cause merging rule to obtain a merged root cause list and the score corresponding to each sub-root cause; and finally outputting a final root cause and optimization suggestion. The method has the advantages of being few in training samples, accurate in positioning, low in computing power consumption, easy to expand, suitable for multi-system networks from 2G to 6G and the like, capable of improving network optimization efficiency and user perception and capable of providing a solution for user-centered network optimization.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of mobile communication and big data technology, and particularly relates to a method and system for locating user communication abnormal root cause by fusing multi-dimensional data. BACKGROUND

[0002] In the existing mobile cellular communication network, network optimization and network planning work is mainly "index-centered" device network type optimization, which means starting from the KPI index of network equipment, and carrying out passive network optimization and network planning work mainly including "equipment hardware fault maintenance and user complaint handling". This network optimization and network planning work is mainly based on network element equipment and focuses on system performance statistics and road testing. Road testing data is based on the test results of one or more professional outdoor mobile phones, and system performance statistics is based on each call, but it only reflects the call process and has no knowledge of the user's complex wireless environment; this is not enough targeted for the wireless quality of a single user, which leads to the fact that network optimization and network planning work is often far away from users and does not play an effective role in serving users and supporting business development. In the important strategic transformation period of enterprises and the development opportunity period of mobile business, transforming the traditional network optimization and network planning work into "user-centered" user type network optimization and network planning can provide high-quality and targeted mobile network service experience for users, and better indicate the planning, optimization, engineering construction and equipment maintenance of mobile cellular communication network, so as to better serve users and support business development. SUMMARY

[0003] The technical purpose of the present application is to solve the above technical problems existing in the current "index-centered" device network type optimization method, and to provide a method for locating user communication abnormal root cause by fusing multi-dimensional data for user communication abnormal call, so as to improve the efficiency of complaint handling and other network optimization and network planning work and improve the perception of complaint users.

[0004] To achieve the above technical purpose, the present application adopts the following technical scheme.

[0005] In a first aspect, the present application provides a method for locating user communication abnormal root cause by fusing multi-dimensional data, comprising: searching for corresponding user call records based on the user ID of the abnormal call; calculating each user call record key data item to obtain user call record key data item sub-root cause and corresponding optimization scheme; calculating the basic network element data item in the user call record to obtain the network element related to the abnormal call; setting abnormal call network element feature data items for the network element, calculating each network element feature data item to obtain each corresponding network element feature data item sub-root cause and corresponding optimization scheme; set weights of each user call record key data item sub-cause and each network element feature data item sub-cause, and merge the sub-causes according to a sub-cause merging rule to obtain a merged cause list, and determine scores corresponding to each sub-cause according to the weights of the sub-causes; output a final root cause of user communication abnormality and an optimization suggestion according to the root cause list and the scores corresponding to each sub-cause in the root cause list.

[0006] In a second aspect, the embodiments of the present application provide a system for locating a root cause of user communication abnormality by fusing multi-dimensional data, comprising: a data acquisition module configured to search for corresponding user call records triggered by a user ID of an abnormal call; a first analysis module configured to calculate each user call record key data item to obtain a user call record key data item sub-cause and a corresponding optimization scheme; a network element locating module configured to calculate basic network element data items in the user call records to obtain abnormal call related network elements; a second analysis module configured to set abnormal call each network element feature data item for the network elements, calculate each network element feature data item, and obtain each corresponding network element feature data item sub-cause and a corresponding optimization scheme; a root cause fusion module configured to set weights of each user call record key data item sub-cause and each network element feature data item sub-cause, and merge the sub-causes according to a sub-cause merging rule to obtain a merged cause list, and determine scores corresponding to each sub-cause according to the weights of the sub-causes; a result output module configured to output a final root cause of user communication abnormality and an optimization suggestion according to the root cause list and the scores corresponding to each sub-cause in the root cause list.

[0007] Compared with the prior art, the method and system for locating a root cause of user communication abnormality by fusing multi-dimensional data provided by the embodiments of the present application have the following beneficial technical effects: the present application proposes a method for locating a root cause of user communication abnormality by fusing multi-dimensional data: fusing multi-dimensional key data in a mobile network, taking user call records as a main line, and triggering core feature data items of multi-dimensional data (such as indexes, parameters, alarms, GIS, parameters, and basic data) related to the call. By setting each type of strongly related data related to the call, multi-dimensional core data related to the abnormal call is searched from a user ID; abnormal call sub-causes are analyzed according to data dimensions, and similar sub-causes are merged; by summarizing and merging, weights are calculated, and finally, sub-causes of user abnormal calls and their priorities are fused to locate a root cause of user communication abnormality. The present application can improve network optimization efficiency and user perception, and provides a solution for network optimization “centered on users”. BRIEF DESCRIPTION OF DRAWINGS

[0008] The drawings described herein are for purposes of illustration only and are not intended to limit the scope of the present disclosure in any way. Additionally, the shapes and proportions of the various components depicted in the drawings are not intended to be limiting, and the dimensions of the various components depicted in the drawings are not intended to be limiting. Those skilled in the art will recognize that various modifications can be made to the embodiments described herein, and that such modifications are intended to be within the scope of the present disclosure. In the drawings: Figure 1 A method flow diagram for fusing multi-dimensional data to locate a user communication abnormality root cause is provided for an embodiment; Figure 2 A method logic design diagram for fusing multi-dimensional data to locate a user communication abnormality root cause is provided for an embodiment; Figure 3 A system structure diagram for fusing multi-dimensional data to locate a user communication abnormality root cause is provided for an embodiment. DETAILED DESCRIPTION

[0009] In order to enable persons skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by persons skilled in the art without creative work should be within the protection scope of the present application.

[0010] The technical solutions in some embodiments of the present application are briefly described as follows: based on user call records and other key data, the call records of abnormal calls can be obtained through user ID (for example, a number) and other information. Key information (for example, release cause) and other information in the call records can obtain some call termination reason judgment information. Through the user call records, information of a cell where the user call is released can be obtained, and then the feature data items (such as alarm, parameter, coverage, interference, and neighboring cell and other multi-aspect data) of each network element of the cell can be fused and analyzed to obtain other root cause judgment information. Based on the rules of weight and influence priority, the information is comprehensively analyzed to obtain the final root cause of the user abnormal call, and corresponding optimization suggestions can be given.

[0011] The present application is further described below in conjunction with the drawings and specific embodiments of the specification.

[0012] An embodiment of the present application provides a method for fusing multi-dimensional data to locate a user communication abnormality root cause, as shown in Figure 1 and Figure 2 The method comprises the following steps: Step S1: Triggered by the user ID of the abnormal call, search for the user call record corresponding to the abnormal call.

[0013] In some embodiments, the user ID commonly used is, for example, the user ID (such as a number), MDN, IMSI, etc.

[0014] The user call record is a detailed file of each call of a user recorded by a mobile communication network device, such as a call detail record (CDR), a measurement report (MR), a communication history report (CHR), and a user billing statement, etc. The user call record corresponding to the abnormal call is obtained by triggering the user ID of the abnormal call.

[0015] Step S2: Calculate the key data items of each user call record to obtain the key data item sub-cause of the user call record and the corresponding optimization scheme.

[0016] The key data items of the user call record and the corresponding judgment threshold are set, the key data items of the user call record are calculated, and if the key data items of the user call record are equal to the corresponding judgment threshold, the key data item sub-cause of the abnormal call is defined, and the corresponding optimization scheme is given.

[0017] For example, the network quality data items in the user call record of the abnormal call are set as the key data items of the user call record, including release cause, signal quality (RSRP (Received Signal Reference Power), RSRQ (Received Signal Reference Quality), SINR (Signal to Interference plus Noise Ratio), TCSI (Traffic Channel Signal Indicator), etc.), service type, abnormal phenomenon category, and neighbor cell switching, etc.

[0018] Step S3: Calculate the basic network element data items in the user call record to obtain the network element related to the abnormal call; set the abnormal call network element characteristic data items for the network element, calculate the network element characteristic data items, obtain the corresponding network element characteristic data item sub-cause and the corresponding optimization scheme.

[0019] In some embodiments, step S3 specifically includes the following steps: Step 3.1: Calculate the basic network element data items in the user call record to obtain the network element related to the abnormal call.

[0020] In some embodiments, the network element related to the user call can be the release network element and the service network element of the abnormal call, for example, including: (1) the network element of abnormal call release, such as base station ID, cell ID, etc; (2) the network element of abnormal call main service, such as base station ID, cell ID, etc; (3) the target network element of abnormal call handover failure, such as base station ID, cell ID, etc.

[0021] Step 3.2 sets the network element characteristic data item of abnormal call, and calculates the sub-root cause and optimization suggestion of each data source (i.e. network element characteristic data item) in stages.

[0022] In some embodiments, the network element characteristic data item can include at least one of network element key data item, index key data item, parameter key data item, neighboring cell key data item, alarm key data item, performance key data item, coverage key data item, and surrounding station key data item.

[0023] In some embodiments, step 3.2 can further include the following steps: Step 3.2.1: calculating the network element key data item to obtain the sub-root cause and the corresponding optimization scheme.

[0024] The network element key data item and the corresponding judgment threshold in the user call record are set manually, the network element key data item is calculated, and if the network element key is equal to the corresponding judgment threshold, it is defined as the sub-root cause of the abnormal call network element characteristic data item, and the corresponding optimization scheme is given.

[0025] For example, the network element key data item includes whether the network element is normally connected to the network or not.

[0026] Step 3.2.2: calculating the index key data item to obtain the sub-root cause and the corresponding optimization scheme.

[0027] The index key data item and the corresponding judgment threshold are set, the index key data item is calculated, and if the index key data item is equal to the judgment value, it is defined as the sub-root cause of the abnormal call network element characteristic data item, and the corresponding optimization scheme is given.

[0028] For example, the index key data item includes the user call release counter.

[0029] Step 3.2.3: calculating the parameter key data item to obtain the sub-root cause and the corresponding optimization scheme.

[0030] The parameter key data item and the corresponding judgment threshold are set, the parameter key data item is calculated, and if the parameter key data item is equal to the judgment threshold, it is defined as the sub-root cause of the abnormal call network element characteristic data item, and the corresponding optimization scheme is given.

[0031] For example, the parameter key data item includes at least one of access parameter, paging parameter, timer parameter, and cell selection / reselection parameter.

[0032] Step 3.2.4: Calculate the neighbor cell key data item to obtain the sub-root cause and the corresponding optimization scheme.

[0033] Set the neighbor cell key data item and the corresponding judgment threshold, calculate the neighbor cell key data item, and if the neighbor cell key data item is equal to the judgment threshold, define it as the abnormal call network element feature data item sub-root cause, and give the corresponding optimization scheme.

[0034] For example, the set neighbor cell key data item includes at least one of the signal strength of the neighbor cell, the inter-cell distance, the handover source cell, the handover target cell, and the received signal strength.

[0035] Step 3.2.5: Calculate the alarm key data item to obtain the sub-root cause and the corresponding optimization scheme.

[0036] Set the alarm key data item and the corresponding judgment threshold, calculate the alarm key data item, and if the alarm key data item is equal to the judgment threshold, define it as the abnormal call network element feature data item sub-root cause, and give the corresponding optimization scheme.

[0037] For example, the set alarm key data item includes at least one of the network element service withdrawal alarm and the connection failure alarm.

[0038] Step 3.2.6: Calculate the performance key data item to obtain the sub-root cause and the corresponding optimization scheme.

[0039] Set the performance key data item and the corresponding judgment threshold, calculate the performance key data item, and if the performance key data item is equal to the judgment threshold, define it as the abnormal call network element feature data item sub-root cause, and give the corresponding optimization scheme.

[0040] For example, the set performance key data item includes at least one of the hidden failure, the RSSI mean value lifting value, the RSSI peak value lifting value, the quiet noise, the peak ROT, and the average ROT.

[0041] Step 3.2.7: Calculate the coverage key data item to obtain the sub-root cause and the corresponding optimization scheme.

[0042] Set the coverage key data item and the judgment threshold, calculate the coverage key data item, and if the coverage key data item is equal to the judgment threshold, define it as the abnormal call network element feature data item sub-root cause, and give the corresponding optimization scheme.

[0043] For example, the set coverage key data item includes at least one of the cell received signal strength, the distance between the NodeB and the UE, the number of sampling points, and the proportion.

[0044] Step 3.2.8: Calculate the surrounding station key data item to obtain the sub-root cause and the corresponding optimization scheme The cell within a certain distance around the abnormal user call record is released as a peripheral cell. Peripheral station key data items and corresponding judgment thresholds are set, the peripheral station key data items are calculated, and if the peripheral station key data items are equal to the judgment thresholds, the abnormal call network element feature data item sub-cause is defined, and the corresponding optimization scheme is given.

[0045] For example, the peripheral station key data items include at least one of interference, PCI mode interference, and engineering construction.

[0046] Step 4: Set the weight of each user call record key data item sub-cause and each network element feature data item sub-cause, and merge the sub-causes according to the similar cause merging rule to obtain a merged root cause list, and determine the score corresponding to each sub-cause according to the weight of each sub-cause.

[0047] Step 4 realizes the similar cause merging of the sub-causes of each data source.

[0048] In the embodiment, step 4 specifically includes the following contents: (1) Set the sub-cause weight: For example, the sub-cause weight can be set according to expert experience as shown in Table 1.

[0049] Table 1: Sub-cause weight setting according to expert experience

[0050] (2) Set the sub-cause merging rule: Each sub-cause outputs a root cause and a score, and the score value is the weight value. The root causes analyzed between the sub-causes and outputted can be merged, and then the scores of the mergable sub-causes are accumulated, and the root causes of the sub-causes are merged together and outputted. The root causes of other sub-causes that cannot be merged are listed separately, and finally a root cause list is formed and arranged in descending order according to the scores of each root cause.

[0051] Using the multidimensional core feature data as the characteristic value, in some embodiments, an AI clustering model (such as an AI clustering model kmeans) can be used to learn and merge each sub-cause.

[0052] Step 5: According to the root cause list and the scores corresponding to each sub-cause therein, output the final root cause of the user communication anomaly and the optimization suggestion.

[0053] Step 5 merges the similar sub-causes to determine the root cause of the fault, and outputs the optimization suggestion. The method for fusing multi-dimensional data to locate the root cause of user communication anomaly provided by the embodiment of the application fuses various key data in a cellular mobile network and performs user-level intelligent optimization. The overall idea of the embodiment is "user ID initiation, domain analysis, and fusion positioning". Various key data in the mobile network are fused, and the user call record is taken as the main line. The core feature data items in the multi-dimensional data such as indicators, parameters, alarms, GIS, parameters, and basic data related to the call are triggered, domain analysis is performed on each sub-feature data (i.e., key data), suspected root causes output by each feature data are combined and weight calculation is performed on the similar root causes, and finally the root cause result and optimization suggestion are fused and positioned, so that the efficiency of network optimization and network planning work such as complaint handling is improved, and the perception of the complaining user is improved.

[0054] The technical solution disclosed in the application can be implemented by using the python programming language, the training framework uses pytorch, and the generative large model technology is used. Based on the current open source large model, not limited to chatglm, Qwen, and the like, business data is collected, and a classification model is constructed.

[0055] In the data calling, evaluation, and analysis links of daily complaint handling, manual operation still occupies a large part. The existing data sources of network optimization have covered user call records, traffic statistical data, parameters, alarms, and basic data. By using the key data items of these data, a series of algorithms are used to automatically associate and intelligently analyze various key data, to determine the root cause of abnormal calls, which can greatly improve the analysis efficiency of user complaints.

[0056] The following embodiment takes an indoor 2G wireless network in a city as an example to provide a method for fusing multi-dimensional data to locate the root cause of user communication anomaly.

[0057] It should be noted that the technical solution of the application has universality for multiple mobile communication systems and is applicable to mobile terminal position positioning in 2G / 3G / 4G / 5G / 5G-A / 6G…… and other mobile communication networks. Therefore, the protection scope of the application should include the field of mobile communication of 2G / 3G / 4G / 5G / 5G-A / 6G…… and the like, and not only the 2G network exemplified in the embodiment.

[0058] According to the complaint information such as the user ID (for example, number) of the abnormal call in the complaint work order, the call record of the complaint call is obtained. The information such as CFC / (Call Failure Cause) CFCQ (Call Failure Cause Quality), TCSI, SM count in the call record can obtain some reason judgment information of the call termination. Through the pilot strength measurement message (Pilot Strength Measurement Message, PSMM) message in the user call record, the information of the user releasing the cell can also be obtained, and then the alarm key data item, parameter key data item, coverage key data item, interference key data item and neighbor cell key data item of the cell can be checked, and some other reason judgment information of the call termination can be obtained. Comprehensive analysis of these information can obtain the final reason of the user complaint, and the corresponding optimization suggestion can be given.

[0059] Considering daily work, the source data of the complaint root cause intelligent optimization in the embodiment is set as the following 6 categories: 2.1. CFC / CFCQ / TCSI judgment sub-function module: The Cfc / CfcQulifer / TCSI value of the user call record is queried, and if it is equal to the judgment threshold, it is defined as a suspected reason.

[0060] The judgment content mainly includes CFC, CFCQ, CFCQ2, service type, abnormal phenomenon category, and suggested checking direction. In addition, whether CFC=2 and CFCQ=109 belong to drop call needs to be judged by referring to TCSI: only under the condition of TCSI Timer=1 and TCSITimer Value=90, CFC=2 and CFCQ=109 are defined as drop call. 2.2. Release Count judgment sub-function module: According to the Last Cell Site Num, Last Sector, Ending Carrier fields of the abnormal user call record, the Release Count of the corresponding base station / sector / carrier is queried, and the query time is the current hour of the abnormal call event. If a Release Count increases, it is defined as a suspected reason.

[0061] The judgment content mainly includes count level, count name, count explanation, service type, and suggested checking direction.

[0062] For the case that the SM count of the abnormal user call record can match the Release Count, the causes of the "CFC / CFCQ / TCSI judgment sub-cause" and the "Release Count judgment sub-cause" can be combined, and for the case that the matching is failed, the causes of the two sub-causes are independently output. The matching principle of the SM count and the Release Count includes the following fields: PCMD Number, SM Count Name, Alloc.Service, main / called, release cause code, count level, count name, checking direction, etc.

[0063] 2.3. Parameter checking sub-function module: The parameters of the release cell in the abnormal user call record are checked whether in the normal range, and if not, the suspected cause is defined. The checked parameters include: (1) handover parameters (cell level); (2) access parameters (cell level); (3) registration parameters (BTS level); If the above parameters are out of the normal range, the abnormal parameters and the values are output, and for the abnormal cell of the handover class parameters, such as the abnormal user call record CFC / CFCQ=2 / 104 (dropped call-handover timeout), the causes of the "CFC / CFCQ / TCSI judgment sub-cause" and the "parameter checking sub-cause" can be combined, and for the case that the matching is failed, the causes of the two sub-causes are independently output.

[0064] 2.4. Neighbor cell checking sub-function module: The oneway / twoway checking and the neighbor cell missing checking are performed according to the neighbor cell of the release cell of the abnormal user call record, and the suspected cause is defined if the oneway / twoway or the neighbor cell missing exists; The oneway / twoway checking algorithm is based on the algorithm in the network optimization platform. The oneway phenomenon abnormal user call record cell should be in the B position, and the twoway phenomenon abnormal user call record cell should be in the B or C position. The neighbor cell missing checking algorithm is based on the homax data.

[0065] For the abnormal user call record release cell with the oneway / twowany or the neighbor cell missing phenomenon, such as the abnormal user call record CFC / CFCQ=2 / 104 (dropped call-handover timeout), the causes of the "CFC / CFCQ / TCSI judgment sub-cause" and the "neighbor cell checking sub-cause" can be combined, and for the case that the matching is failed, the causes of the two sub-causes are independently output.

[0066] 2.5. Key alarm checking sub-function module: According to the abnormal user call record release cell checking base station key alarm and implicit failure, checking the existence of alarm is defined as suspected reason; The base station key alarm includes: (1) TFU alarm, base station alarm, 1x CRC alarm, DS1 alarm, CCU alarm, CCU activation failure alarm, CCU NO HEARTBEAT alarm; (2) CBR / TXAMP / RRH alarm; The base station implicit failure includes: (1) CCU implicit failure judgment (whether the drop call is concentrated on the same CCU); (2) Trunk implicit failure, including Speech Handler trunk group and Speech Handler trunk; For the abnormal user call record release cell with TFU alarm, if the CFC / CFCQ of the abnormal user call record is 2 / 104 (drop call - handover timeout), the reasons of "CFC / CFCQ / TCSI judgment sub-cause" and "key alarm checking sub-cause" can be combined. For the case that cannot be matched, the reasons of the two sub-causes are independently output.

[0067] 2.6. RSSI checking sub-function module; According to the abnormal user call record release cell, the RSSI index of the current 1 hour is checked (threshold adjustable), and if it exceeds the threshold, it is defined as suspected reason.

[0068] 2.7. Coverage judgment sub-function module: According to the abnormal user call record release cell, the judgment of over / weak coverage and pilot pollution is made, and the judgment algorithm is based on the network optimization platform algorithm. If there is over / weak coverage or pilot pollution in this place, it is defined as suspected reason.

[0069] In addition, the GIS positioning of the abnormal user call record can also be performed, and the weak coverage area layer found by daily optimization can be called to judge whether the abnormal user call record is located in the weak coverage area. If the GIS positioning of the abnormal user call record is located in the weak coverage area, the "weak coverage" reason item is output.

[0070] 2.8. Peripheral station analysis sub-function module: According to the abnormal user call record release cell and the cells within a certain distance (threshold adjustable) around the complaint point, the base stations in the complaint period (threshold adjustable) are checked, and if there is an abnormality, it is defined as suspected reason.

[0071] Based on the method of the patent, 5G / 4G / 2G user wireless complaint root cause positioning (hereinafter referred to as "intelligent analysis") has been applied in network optimization platform construction and application. The following is a case analysis of a live network. One complaint work order received by the network optimization platform on a certain day was selected for intelligent analysis, and the analysis results were fast and accurate. The final analysis report is as follows: 5G / 4G / 2G wireless complaint intelligent analysis report: 1. Intelligent analysis summary: Work order number: 10000; User number: 189××××××××; Service type: all; Problem occurrence time: 2024-05-06 22:00:00.0; Problem phenomenon category: all; Problem occurrence address longitude: 0; Problem occurrence address latitude: 0; Main / called: all; Suspected fault cause explanation, as shown in Table 2.

[0072] Table 2 Suspected fault cause explanation

[0073] 2. Sub-root cause analysis results: 2.1 Cfc / CfcQUalifer / TCSI judgment module: (1) Analysis data as shown in Table 3.

[0074] Table 3 Cfc / CfcQUalifer / TCSI judgment module analysis data

[0075] (2) Analysis results: poor reverse link quality causes call drop.

[0076] 2.2 Release Count judgment module: (1) Analysis data as shown in Table 4.

[0077] Table 4 Release Count judgment module analysis data

[0078] (2) Analysis results: 3G caller TCC failure.

[0079] 2.3 Parameter checking module: (1) Analysis data as shown in Table 5.

[0080] Table 5 Parameter checking module analysis data

[0081] (2) Analysis result: No abnormal parameter.

[0082] 2.4, Neighbor verification module: (1) Analysis data: (a) Neighbor missing verification as shown in Table 6.

[0083] Table 6 Neighbor missing verification analysis table

[0084] (b) Neighbor oneway verification: No neighbor oneway phenomenon.

[0085] (c) Neighbor twoway verification as shown in Table 7.

[0086] Table 7 Neighbor twoway verification analysis data

[0087] (2) Analysis result: There is neighbor missing, and there is neighbor TWOWAY.

[0088] 2.5, Key alarm verification module: (1) Analysis data: (a) Key alarm verification: No key alarm.

[0089] (b) CCU implicit failure verification as shown in Table 8.

[0090] Table 8 CCU implicit failure verification analysis data

[0091] (c) Implicit failure verification: No Trunk implicit failure.

[0092] (2) Analysis result: There is CCU implicit failure.

[0093] 2.6, RSSI verification module: (1) Analysis data as shown in Table 9.

[0094] Table 9 RSSI verification module analysis data

[0095] (2) Analysis result: No interference.

[0096] 2.7, Coverage judgment module: (1) Analysis data: (a) Weak coverage verification as shown in Table 10.

[0097] Table 10 Weak coverage verification analysis data

[0098] (b) Overlap check as Table 11.

[0099] Table 11 Overlap check analysis data

[0100] (c) Pilot pollution check as Table 12.

[0101] Table 12 Pilot pollution check analysis data

[0102] (2) Analysis result: Release cell "××××" exists over coverage, and release cell "××××" exists pilot pollution.

[0103] 2.8, Peripheral station analysis module: (1) Analysis data (omitted in the present application); (2) Analysis result: There is no abnormality in the peripheral station.

[0104] 2.9, Expert opinion or historical complaint library module: (1) Analysis data (omitted in the present application); The expert does not give guiding suggestions (omitted in the present application).

[0105] The embodiment realizes the automation and intelligentization of 5G / 4G / 2G wireless complaint analysis, and provides engineers with timely, accurate and intuitive analysis results and optimization suggestions, so as to ensure that engineers can quickly solve complaints and improve user perception.

[0106] Based on the same inventive concept as the above embodiment, the embodiment of the present application also provides a system for fusing multi-dimensional data to locate the root cause of user communication anomaly, as shown in Figure 3 , comprising a data acquisition module, a first analysis module, a network element positioning module, a second analysis module, a root cause fusion module and a result output module.

[0107] The data acquisition module is used to search for corresponding user call records with the user ID of the abnormal call as a trigger.

[0108] The first analysis module is used to calculate each user call record key data item to obtain a user call record key data item sub-root cause and a corresponding optimization scheme.

[0109] The network element positioning module is used to calculate the basic network element data item in the user call record to obtain the network element related to the abnormal call.

[0110] The second analysis module is configured to set each network element feature data item for the abnormal call of the network element, calculate each network element feature data item, obtain each corresponding network element feature data item sub-root cause and a corresponding optimization scheme.

[0111] The root cause fusion module is configured to set weights of each user call record key data item sub-root cause and each network element feature data item sub-root cause, merge the sub-root causes according to a sub-root cause merging rule to obtain a merged root cause list, and determine scores corresponding to each sub-root cause according to the weights of the sub-root causes.

[0112] The result output module is configured to output a user communication abnormal final root cause and an optimization suggestion according to the root cause list and the scores corresponding to each sub-root cause in the root cause list.

[0113] The system is applicable to 2G, 3G, 4G, 5G, 5G-A and 6G mobile communication networks.

[0114] The embodiment realizes automation and intelligentization of 5G / 4G / 2G wireless complaint analysis, provides engineers with timely, accurate and intuitive analysis results and optimization suggestions, ensures engineers to quickly solve complaints, and improves user perception.

[0115] The above describes in detail a method and system for fusing multi-dimensional data to locate a user communication abnormal root cause provided by the present application, specific examples are applied to explain the principle and implementation mode of the present application, and the above embodiment is only used to help understand the concept of the present application and should not be understood as a limitation on the protection scope of the present application.

Claims

1. A method of fusing multi-dimensional data to locate a root cause of a user communication anomaly, the method comprising: The method comprises: triggering the search of the corresponding user call record with the user ID of the abnormal call; calculating each user call record key data item to obtain a user call record key data item sub-root cause and a corresponding optimization scheme; calculating the basic network element data item in the user call record to obtain an abnormal call related network element; setting an abnormal call network element feature data item for the network element, calculating each network element feature data item to obtain each corresponding network element feature data item sub-root cause and a corresponding optimization scheme; setting the weight of each user call record key data item sub-root cause and each network element feature data item sub-root cause, and merging the sub-root causes according to the sub-root cause merging rule to obtain a merged root cause list, and determining the score corresponding to each sub-root cause according to the weight of each sub-root cause; outputting the final root cause of the user communication abnormality and the optimization suggestion according to the root cause list and the score corresponding to each sub-root cause in the root cause list.

2. The method of claim 1, wherein the method further comprises: The user call record key data item comprises at least one of the user call release reason, the signal quality, the service type, the abnormal phenomenon category, and the adjacent cell handover. If the user call record key data item is equal to the corresponding preset judgment threshold, it is defined as a user call record key data item sub-root cause, and a corresponding optimization scheme is generated.

3. The method of claim 1, wherein the method further comprises: The network element feature data item comprises an index key data item, and the index key data item comprises a user call release frequency; the index key data item and the corresponding judgment threshold are set, and if the index key data item is equal to the corresponding judgment threshold, it is defined as a network element feature data item sub-root cause, and a corresponding optimization scheme is generated.

4. The method of claim 1, wherein the method further comprises: The network element feature data item comprises a parameter key data item, and the parameter key data item comprises at least one of the access parameter, the paging parameter, the timer parameter, and the cell selection / reselection parameter; the parameter key data item and the corresponding judgment threshold are set, and if the parameter key data item is equal to the corresponding judgment threshold, it is defined as a network element feature data item sub-root cause, and a corresponding optimization scheme is generated.

5. The method of claim 1, wherein the method further comprises: The network element feature data item comprises a neighboring cell key data item, and the neighboring cell key data item comprises at least one of the adjacent cell signal strength, the inter-cell distance, the handover source cell, the handover target cell, and the received signal strength; the neighboring cell key data item and the corresponding judgment threshold are set, and if the neighboring cell key data item is equal to the corresponding judgment threshold, it is defined as a network element feature data item sub-root cause, and a corresponding optimization scheme is generated.

6. The method of claim 1, wherein the method further comprises: The network element feature data item comprises an alarm key data item, and the alarm key data item comprises at least one of the network element service withdrawal alarm and the connection failure alarm; if the alarm key data item is equal to the corresponding judgment threshold, it is defined as a network element feature data item sub-root cause, and a corresponding optimization scheme is generated.

7. The method of claim 1, wherein the method further comprises: The network element feature data item comprises at least one of the network element key data item, the index key data item, the parameter key data item, the neighboring cell key data item, the alarm key data item, the performance key data item, the coverage key data item, and the surrounding station key data item.

8. The method of claim 1, wherein the method further comprises: The method further comprises: performing similarity analysis on the root causes output by each sub-root cause, accumulating the weight values of the sub-root causes that can be merged, and merging the root cause outputs; separately listing the sub-root causes that cannot be merged, and arranging them in descending order of weight values.

9. The method of claim 1, wherein the method further comprises: The AI clustering model includes a kmeans model.

10. A system for fusing multi-dimensional data to locate root cause of user communication anomalies, characterized in that, The method comprises: The data acquisition module is configured to search for corresponding user call records based on a user ID of an abnormal call as a trigger; The first analysis module is configured to calculate each user call record key data item to obtain user call record key data item sub-root causes and corresponding optimization schemes; The network element positioning module is configured to calculate basic network element data items in the user call records to obtain network elements related to the abnormal call; The second analysis module is configured to set abnormal call network element feature data items for the network elements, calculate each network element feature data item, obtain each corresponding network element feature data item sub-root cause, and obtain corresponding optimization schemes; The root cause fusion module is configured to set weights of each user call record key data item sub-root cause and each network element feature data item sub-root cause, merge the sub-root causes according to a sub-root cause merging rule to obtain a merged root cause list, and determine scores corresponding to each sub-root cause according to the weights of the sub-root causes; The result output module is configured to output user communication abnormal final root causes and optimization suggestions based on the root cause list and the scores corresponding to each sub-root cause in the root cause list.

11. The system of claim 10, wherein, The system is applicable to 2G, 3G, 4G, 5G, 5G-A, and 6G mobile communication networks. ​

Images