Network poor quality analysis method and device, electronic equipment and computer readable medium

Abstract

The invention provides a network poor quality analysis method and device, electronic equipment and a computer readable medium. The method comprises the following steps: acquiring application record information XDR and wireless measurement report information MR of a communication network to be analyzed; matching the XDR with the MR, and combining network quality information and position information in the MR into the XDR matched with the MR to obtain first associated record information; and according to the first association record information, analyzing the network quality. According to the embodiment of the invention, the effect of network poor quality analysis can be improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technologies, and particularly to a method, an apparatus, an electronic device, and a computer-readable medium for analyzing poor network quality. Background Art

[0002] Providing a satisfactory user network experience is one of the important responsibilities of network operators. This not only helps to improve customer loyalty, but also attracts new customers, maintains the brand reputation, and ensures the sustainable growth of the business. Therefore, starting from the user perception, accurately identifying network quality problems and guiding network optimization are important means to improve the operation and maintenance efficiency of mobile communication networks and improve network quality.

[0003] Among them, a network black spot refers to an area in a wireless communication network where the signal coverage is poor, the signal is weak, or the interference is strong, which affects the user's communication and data experience. In the related art, the root cause of poor service quality (referred to as the root cause of poor quality) is usually determined based on manual experience, and it is impossible to perform intelligent analysis on network black spots, making it difficult to effectively support network optimization. Summary of the Invention

[0004] The present disclosure provides a method, an apparatus, an electronic device, and a computer-readable medium for analyzing poor network quality.

[0005] In a first aspect, an embodiment of the present disclosure provides a method for analyzing poor network quality, the method including:

[0006] Obtaining application record information XDR and radio measurement report information MR of a communication network to be analyzed; matching the XDR and the MR, and merging the network quality information and location information in the MR into the XDR that matches the MR to obtain first associated record information; analyzing poor network quality according to the first associated record information.

[0007] In a second aspect, an embodiment of the present disclosure provides an apparatus for analyzing poor network quality, the apparatus including:

[0008] An information acquisition module, configured to obtain application record information XDR and radio measurement report information MR of a communication network to be analyzed;

[0009] An information matching module, configured to match the XDR and the MR, and merge the network quality information and location information in the MR into the XDR that matches the MR to obtain first associated record information;

[0010] A poor quality analysis module, configured to analyze poor network quality according to the first associated record information.

[0011] In a third aspect, an embodiment of the present disclosure provides an electronic device, which includes a memory and a processor; the memory stores a computer program executable by the processor, and when the computer program is executed by the processor, the above-mentioned network quality degradation analysis method is implemented.

[0012] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned network quality degradation analysis method is implemented.

[0013] According to the embodiments of the present disclosure, it is possible to obtain application record information XDR and radio measurement report information MR of a communication network to be analyzed; match XDR and MR, and merge the network quality information and location information in MR into the XDR that matches MR to obtain associated record information; analyze network quality degradation according to the associated record information, thereby improving the analysis accuracy and analysis effect of network quality degradation analysis and enhancing the operation and maintenance efficiency.

[0014] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The drawings are used to provide a further understanding of the present disclosure, and constitute a part of the specification. They are used together with the embodiments of the present disclosure to explain the present disclosure, and do not constitute a limitation to the present disclosure. By describing the detailed exemplary embodiments with reference to the drawings, the above and other features and advantages will become more obvious to those skilled in the art. In the drawings of the embodiments of the present disclosure:

[0016] Figure 1 is a flowchart of a network quality degradation analysis method provided by an embodiment of the present disclosure;

[0017] Figure 2 is a schematic diagram of an index quality degradation status field provided by an embodiment of the present disclosure;

[0018] Figure 3 is a schematic diagram of quality degradation area clustering provided by an embodiment of the present disclosure;

[0019] Figure 4 is a schematic diagram of quality degradation threshold learning provided by an embodiment of the present disclosure;

[0020] Figure 5 is a schematic diagram of the relationship between network quality information and network service quality indicators provided by an embodiment of the present disclosure;

[0021] Figure 6 is a block diagram of the composition of a network quality degradation analysis device provided by an embodiment of the present disclosure;

[0022] Figure 7 A block diagram of an electronic device provided by an embodiment of the present disclosure. Specific embodiments

[0023] To enable those skilled in the art to better understand the technical solutions of the present disclosure, the network quality degradation analysis method provided by the embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings.

[0024] The present disclosure will be described more fully hereinafter with reference to the accompanying drawings. However, the illustrated embodiments may be embodied in different forms and the present disclosure should not be construed as limited to the embodiments set forth below. On the contrary, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the present disclosure to those skilled in the art.

[0025] The accompanying drawings of the embodiments of the present disclosure are used to provide a further understanding of the embodiments of the present disclosure, and constitute a part of the specification. Together with the detailed embodiments, they are used to explain the present disclosure and do not constitute a limitation to the present disclosure. By describing the detailed embodiments with reference to the accompanying drawings, the above and other features and advantages will become more apparent to those skilled in the art.

[0026] The present disclosure may be described with reference to the plan views and / or cross-sectional views by means of the ideal schematic diagrams of the present disclosure. Therefore, the example illustrations can be modified according to the manufacturing technology and / or tolerances.

[0027] In the case of no conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other.

[0028] The terms used in the present disclosure are only for describing specific embodiments and are not intended to limit the present disclosure. As used in the present disclosure, the term "and / or" includes any and all combinations of one or more related listed items. As used in the present disclosure, the singular forms "a" and "the" are also intended to include the plural forms unless the context clearly indicates otherwise. As used in the present disclosure, the terms "include", "made of...", specify the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their groups.

[0029] Unless otherwise defined, all terms (including technical and scientific terms) used in the present disclosure have the same meaning as commonly understood by those of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and the present disclosure, and will not be interpreted as having an idealized or overly formal meaning unless the present disclosure clearly defines so.

[0030] The present disclosure is not limited to the embodiments shown in the drawings, but includes modifications to the configurations formed based on the manufacturing process. Therefore, the regions illustrated in the drawings have schematic properties, and the shapes of the regions shown in the figures illustrate the specific shapes of the regions of the components, but are not intended to be restrictive.

[0031] Network operators have carried out a series of special actions closely around their main responsibilities and main businesses, deeply explored customer needs, focused on the key links of users' usage experience, investigated the potential risks affecting users' network experience, helped the general public enjoy high-speed networks, continuously strengthened the network-service coordination ability, and made it possible to respond immediately to users' demands.

[0032] Among them, a network black spot refers to an area in a wireless communication network where the signal coverage is poor, the signal is weak, or the interference is strong, affecting users' communication and data experience.

[0033] In related technologies, usually, the root cause of poor service quality (referred to as the root cause of quality degradation) is determined based on manual experience, and it is impossible to perform intelligent analysis on network black spots, making it difficult to effectively support network optimization. Moreover, only service quality indicators and network quality indicators are provided, but there is a lack of comprehensive evaluation of the economic value of network black spots, which makes it difficult for operation and maintenance personnel to accurately judge the impact of network black spots on services and the economic value of network optimization. In addition, for the massive data correlation of wireless measurement reports MR (Measurement Report) and service XDR (x(application)detail record, detailed record of any application) based on DPI (Deep Packet Inspection), there is a lack of a flexible correlation screening strategy, which requires consuming a large amount of storage and computing resources, will increase the network operation and maintenance cost, and is difficult to promote. The disadvantages of related technologies result in problems such as low operation and maintenance efficiency due to the lack of intelligent analysis of the root cause of network black spots, difficulty in evaluating the economic value of network black spots, and increased investment costs due to the lack of a flexible massive data correlation strategy in the actual application of network black spots.

[0034] According to the embodiments of the present disclosure, a flexible correlation screening strategy is set to reduce the amount of data to be processed, which can save storage and computing resources; through intelligent analysis to identify quality degradation regions (such as black spots), perform root cause analysis of black spots and evaluate the value of black spots, etc., it can enable the identification and analysis of network black spots to be truly applied to production, improve operation and maintenance efficiency, shorten the fault recovery cycle, and enhance user satisfaction.

[0035] The network quality degradation analysis method according to an embodiment of the present disclosure can be executed by an electronic device such as a terminal device or a server. The terminal device can be a user equipment (UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (PDA), a handheld device, a computing device, a vehicle-mounted device, a wearable device, etc. The server can be an independent physical server, a server cluster composed of multiple physical servers, or a cloud server capable of performing cloud computing. This method can be implemented by a processor invoking computer-readable program instructions stored in a memory.

[0036] Figure 1 FIG. is a flowchart of a network quality degradation analysis method provided by an embodiment of the present disclosure. As Figure 1 shown, the network quality degradation analysis method according to an embodiment of the present disclosure includes:

[0037] Step S11, obtaining application record information XDR and radio measurement report information MR of a communication network to be analyzed;

[0038] Step S12, matching the XDR and the MR, and merging the network quality information and location information in the MR into the XDR that matches the MR to obtain first associated record information;

[0039] Step S13, analyzing network quality degradation according to the first associated record information.

[0040] For example, the communication network to be analyzed can be a wireless communication network (or a mobile communication network) covering a preset geographical location. The preset geographical location can be a geographical area or an administrative area, such as a district or county administrative area; it can also be a custom area, such as grouping wireless cells according to scenarios. Among them, a wireless cell is, for example, the radiation area or radiation sector of a network element of a network device (base station). The radiation area of a base station may be divided into 3-4 wireless cells, and the wireless cell may cover an actual residential community, building, street, etc. It should be understood that those skilled in the art can set the specific range of the preset geographical location according to actual situations, and the present disclosure does not limit this.

[0041] In some possible implementation manners, the application record information XDR (x (application) detail record, detailed record of any application) may be the service XDR information obtained by identifying and acquiring based on the DPI technology (which may also be referred to as service signaling process data). When a user uses an application in a terminal (such as a smart phone), for example, during instant messaging, payment, web browsing, or video playing, application record information XDR will be generated. Each piece of application record information XDR may include service type (such as video playing), service start time, service duration, data traffic used, data buffering time, whether there is lag, network service quality metrics, etc.

[0042] In some possible implementation manners, the wireless measurement report information MR (Measurement Report) is the information reported by the terminal side to the base station. The terminal side may report periodically (in seconds, such as 5s, 10s) and report events (such as switching from a 4G network to a 5G network). Each piece of wireless measurement report information MR may include the wireless network signal condition, the location of the terminal itself (such as the identifier of the wireless cell where it is located), the terminal identifier ID, the occurrence time of the network service, network quality information, etc.

[0043] In some possible implementation manners, both the application record information XDR and the wireless measurement report information MR are information collected by an external system, and the information collected by the external system may further include network element network management performance data (including wireless alarm data, etc.). Among them, the electronic device (such as a cluster software platform capable of distributed computing and storage) executing the embodiments of the present disclosure is docked with the external system based on a general interface, requiring network reachability between the two and the bandwidth to meet the transmission requirements. The present disclosure does not limit the manner of collecting information by the external system, the specific types and formats of the collected information, and the specific contents included in the collected application record information and wireless measurement report information, etc.

[0044] In some possible implementation manners, the application record information XDR and the wireless measurement report information MR of the communication network to be analyzed within a preset historical time period may be obtained in step S11. Among them, the historical time period may be a period of historical time to be analyzed, such as the past 1 day, 1 week, or 1 month, etc. And, a busy time range T busy may also be set, that is, the busy time interval within a day, such as [7, 23], indicating that 7 o'clock to 23 o'clock (including 23 o'clock) is the busy time. In this way, the amount of data to be analyzed can be reduced and the calculation efficiency can be improved.

[0045] In some possible implementation manners, the set P of wireless cells corresponding to a preset geographical location may be determined according to the wireless cell engineering parameter configuration and the hotspot area configuration. cell; Determine the key service P according to the analysis scenario app , for popular or concerned services, a service is uniquely determined by the service major category identifier and the service minor category identifier. For example, (10, 100), where 10 represents the service major category identifier and 100 represents the service minor category identifier; determine the busy hour time range T within the historical time period according to the preset busy hour time range busy .

[0046] In some possible implementation manners, based on the wireless cell set P cell 、the key service P app 、the busy hour time range T busy , screen and filter the received application record information to obtain the application record information XDR to be matched; based on the wireless cell set P cell 、the busy hour time range T busy , screen and filter the received wireless measurement report information to obtain the wireless measurement report information MR to be matched. In this way, the amount of data to be matched can be reduced, and the amount of stored data and the amount of calculation can be reduced.

[0047] In some possible implementation manners, the application record information to be matched and the wireless measurement report information can be stored in different folders respectively, and partitioned according to the date and the hash value of the wireless cell. The date here is the timestamp after rounding based on the service occurrence time T proc according to the preset time window (associated time granularity Gr).

[0048] In step S12, match the XDR and the MR, and merge the network quality information and the location information in the MR into the XDR that matches the MR to obtain the first associated record information.

[0049] In some possible implementation manners, an associated time granularity can be set, and the value range is, for example, 15 minutes, 30 minutes, 1 hour, 1 day, etc. In this way, the historical time period can be divided into multiple associated time granularities, and the application record information and the wireless measurement report information can be matched, thereby reducing the amount of calculation for matching.

[0050] In some possible implementation manners, for any associated time granularity (for example, from 10:00 to 11:00 in the morning of a certain day), the application record information and the radio measurement report information within the associated time granularity can be matched. Among them, preliminary matching can be performed through the user identifier, radio cell identifier, and associated time difference in the application record information and the radio measurement report information. For example, for any piece of application record information, the radio measurement report information with the same user identifier, the same radio cell identifier, and the associated time difference less than or equal to the associated time window can be determined as the alternative radio measurement report information. The associated time difference can be jointly determined by the first service occurrence time and the service duration in the application record information and the second service occurrence time in the radio measurement report information.

[0051] In some possible implementation manners, after preliminary matching, one piece of application record information may match multiple alternative radio measurement report information. First, according to the radio measurement report information positioning type (such as GPS positioning, server positioning (such as triangulation), etc.), the radio measurement report information with a higher positioning type accuracy can be preferentially selected; if the positioning types are the same, then the radio measurement report information with a smaller time difference can be preferentially selected, and finally, one piece of radio measurement report information that matches the application record information can be determined. Furthermore, the network quality information and location information in the radio measurement report information MR can be merged into the application record information to obtain the first associated record information, which can also be called WQ-XDR (XDR with Wireless quality data, XDR carrying wireless network quality data).

[0052] In this way, by respectively matching all the application record information in the associated time granularity and adding the network quality information and location information to the application record information XDR of the matched radio measurement report information, multiple first associated record information of the associated time granularity can be obtained. By respectively performing matching processing on multiple associated time granularities, the first associated record information of the historical time period can be obtained.

[0053] In step S13, according to the first associated record information, the network quality degradation is analyzed. Among them, the network quality degradation analysis can include the identification of the degraded area, the root cause analysis of the degraded area, the value level analysis of the degraded area, etc.

[0054] For the identification of poor-quality regions, multiple network service types can be distinguished according to service characteristics and content types used in the application, such as short videos, long videos, live broadcasts, page views, games, etc., and network service quality indicators for each network service can be set, such as download rate indicators, buffering time indicators, etc. For each indicator, a corresponding composite indicator formula can be set in combination with network quality information that can delimit radio-side problems (for example, downlink RTT delay, if this delay is large, it is considered a problem on the wireless network side), and an indicator poor-quality threshold can be set.

[0055] In some possible implementation manners, the first associated record information can be processed item by item. For any piece of first associated record information, the corresponding network service type and the corresponding network service quality indicator can be determined, the indicator value can be calculated using the corresponding composite indicator formula, and then compared with the indicator poor-quality threshold to identify the poor-quality state of the indicator, and each indicator poor-quality state can be recorded into the indicator poor-quality state field (S pq variable) by 1 bit. After the poor-quality states of all indicators of the first associated record information are processed, the indicator poor-quality state field (S pq value) is added to the first associated record information to obtain the second associated record information.

[0056] In some possible implementation manners, the second associated record information can be aggregated according to dimensions such as user information, region information, network service type information, location information, etc. For example, it can be aggregated once by service category, and then aggregated a second time by location information (for example, grid location, each grid is 50m * 50m), and the poor-quality ratio of the grid (i.e., the poor-quality ratio) can be calculated for the indicators of each network service type; if the poor-quality ratio is greater than or equal to the regional poor-quality threshold, the corresponding grid can be determined as one or more poor-quality grids of the network service type. In this way, by processing multiple grids of the preset geographical location respectively, the poor-quality grids of the network service quality can be determined.

[0057] In some possible implementation manners, the poor-quality grids in the preset geographical location can be clustered, adjacent poor-quality grids can be identified as poor-quality regions (such as network black spots), and a poor-quality region identifier can be set for the poor-quality regions, thereby completing the process of identifying the poor-quality regions. Among them, the poor-quality region can refer to a network black spot, that is, a region with poor signal coverage, weak signal or strong interference that appears in the wireless communication network; it can also be other types of poor-quality regions, and the present disclosure does not limit this.

[0058] In some possible implementation manners, after obtaining the poor-quality regions, root cause analysis of the poor quality, value level analysis, etc. can be performed on each poor-quality region to further improve the effect of network poor-quality analysis.

[0059] According to an embodiment of the present disclosure, application record information XDR and radio measurement report information MR of a communication network to be analyzed can be obtained; the XDR and the MR are matched, and the network quality information and location information in the MR are merged into the XDR matched with the MR to obtain associated record information; and the network quality degradation is analyzed according to the associated record information, so as to improve the analysis accuracy and analysis effect of the network quality degradation analysis and improve the operation and maintenance efficiency.

[0060] The network quality degradation analysis method according to an embodiment of the present disclosure will be described in detail below.

[0061] As described above, the application record information XDR and the radio measurement report information MR of the communication network to be analyzed can be obtained in step S11. Among them, the communication network to be analyzed may be a wireless communication network (or a mobile communication network) covering a preset geographical location, and the preset geographical location P zone may be an administrative region or a custom region. The administrative region is, for example, a district or county-level administrative region, and the custom region is, for example, a region corresponding to a plurality of wireless cells grouped according to a scenario. The preset geographical location P zone may be represented as, for example, (0, 200), where 0 represents the region type and 200 represents the region number. The specific region type and region number are defined according to the actual application scenario, and the present disclosure does not limit this.

[0062] In some possible implementation manners, each piece of application record information XDR may include a network service type (such as video playback), a service start time, a service duration, used data traffic, a data buffering time, whether there is a freeze, a network service quality index, etc. Each piece of radio measurement report information MR includes a wireless network signal condition, the location of the terminal itself (such as the identifier of the wireless cell where it is located), a terminal identifier ID, the occurrence time of the network service, network quality information, etc.

[0063] In some possible implementation manners, the XDR and the MR may be matched in step S12.

[0064] Among them, matching the XDR and the MR includes matching the XDR and the MR according to at least one of the following association strategies or a combination thereof:

[0065] A time association strategy, which matches the XDR and the MR based on a preset time strategy;

[0066] A region association strategy, which matches the XDR and the MR based on a preset geographical location;

[0067] A service type association strategy, which matches the XDR and the MR based on a preset service type.

[0068] That is to say, at least one of the time correlation strategy, region correlation strategy, and service type correlation strategy can be adopted to match XDR and MR, so as to reduce the amount of data to be matched and improve the accuracy of matching.

[0069] Table 1 shows the parameters that may be used during information acquisition and information matching and the corresponding parameter descriptions, which may include the correlation time granularity, the XDR delay threshold for application record information, the MR delay threshold for radio measurement report information, key services, preset geographical locations, busy time ranges, correlation time windows, etc.

[0070] Table 1

[0071]

[0072] In the time correlation strategy, data in the historical time period can be matched. Moreover, the busy time range T busy can also be set, that is, the busy time interval within a day, such as [7, 23], indicating that 7 o'clock to 23 o'clock (including 23 o'clock) is the busy time. In this way, the amount of data to be analyzed can be reduced and the calculation efficiency can be improved.

[0073] In the region correlation strategy, the preset geographical location can be determined. The set of radio cells P corresponding to the preset geographical location can be determined according to the radio cell engineering parameters (such as frequency, location and other engineering parameters) configuration and the hot spot area configuration. cell .

[0074] In the region correlation strategy, the preset service type can be determined. According to the analysis scenario, the key service P app can be determined, that is, the popular or concerned service. A service is uniquely determined by the service major category identifier and the service minor category identifier. For example, (10, 100), where 10 represents the service major category identifier and 100 represents the service minor category identifier.

[0075] Furthermore, based on the set of radio cells P cell , the key service P app , and the busy time range T busy , the received application record information is screened and filtered to obtain the application record information to be matched; based on the set of radio cells P cell , and the busy time range T busy , the received radio measurement report information is screened and filtered to obtain the radio measurement report information to be matched. In this way, based on flexible screening strategies such as the region range, time parameters, and key services, the on-demand matching of application record information and radio measurement report information is realized, saving a large amount of storage and computing resources. At the same time, the screening strategy can be dynamically adjusted, improving the flexibility of processing.

[0076] In some possible implementation manners, the application record information to be matched and the radio measurement report information may be stored in different folders respectively, and partitioned according to the date and the hash value of the radio cell. The date herein is based on the service occurrence time T proc The time stamp after being rounded according to a preset time window (associated time granularity Gr).

[0077] In some possible implementation manners, in the time association policy, the corresponding time policy may include at least one of the following: the associated time granularity, used to characterize the time interval corresponding to the XDR and the MR participating in the matching;

[0078] The XDR delay threshold, used to characterize the maximum delay of the reporting time of the XDR relative to the time interval;

[0079] The MR delay threshold, used to characterize the maximum delay of the reporting time of the MR relative to the time interval;

[0080] The busy time range, used to characterize the busy time range of the communication network in a day;

[0081] The associated time window, used to characterize the time difference range between the matching XDR and MR.

[0082] For example, the application record information XDR and the radio measurement report information MR may be matched according to multiple associated time granularities. Among them, the associated time granularity G r has a value range such as 15 minutes, 30 minutes, 1 hour, 1 day, etc., so that the calculation amount of each association can be reduced and the calculation accuracy can be improved.

[0083] In some possible implementation manners, a turning-over policy of the application record information and the radio measurement report information within the associated time granularity may be set, that is, within the time interval of the associated time granularity, every time an application record information is obtained, the time T of the last received application record information in this time interval will be registered last , and when the difference between the current time and T last is greater than or equal to the XDR delay threshold D xdr , that is, when the maximum delay of the reporting time of the XDR relative to this time interval is reached, it is considered that the application record information in this time interval has been completely received, and the application record information is turned over; similarly, within this time interval, every time a radio measurement report information is obtained, the time T of the last received radio measurement report information in this time interval will be registered last , and when the difference between the current time and T last is greater than or equal to the MR delay threshold D mrWhen the reporting time of MR reaches the maximum time delay with respect to this time interval, it is considered that the wireless measurement report information for this time interval has been completely received, and the wireless measurement report information is turned over.

[0084] In this case, for any time interval corresponding to an associated time granularity (for example, from 10:00 am to 11:00 am on a certain day), the above-mentioned turning-over strategy can be adopted to determine multiple application record information and multiple wireless measurement report information within this time interval, and then match the two.

[0085] In some possible implementation manners, preliminary matching can be performed through the user identifier, wireless cell identifier, and associated time difference in the application record information and the wireless measurement report information. For example, for any piece of application record information, the wireless measurement report information with the same user identifier, the same wireless cell identifier, and the associated time difference less than or equal to the associated time window can be determined as the alternative wireless measurement report information. The associated time window is used to represent the time difference range between the matching XDR and MR.

[0086] Among them, the associated time difference can be jointly determined by the first service occurrence time and service duration in the application record information and the second service occurrence time in the wireless measurement report information. As shown in formula (1):

[0087] Associated time difference = |XDR.T proc + XDR.duration / 2 - MR.T proc | ≤ W a (1)

[0088] In formula (1), XDR.T proc represents the first service occurrence time in the application record information; XDR.duration represents the service duration; MR.T proc represents the second service occurrence time in the wireless measurement report information; W a represents the associated time window.

[0089] In some possible implementation manners, after preliminary matching, one piece of application record information may match multiple alternative wireless measurement report information. First, according to the wireless measurement report information positioning type (such as GPS positioning, server positioning (such as triangulation), etc.), the wireless measurement report information with a higher positioning type accuracy (such as GPS positioning) can be preferentially selected; if the positioning types are the same, the wireless measurement report information with a smaller associated time difference can be preferentially selected, and finally, one piece of wireless measurement report information that matches this application record information can be determined.

[0090] In this way, by separately matching all the application record information in this associated time granularity, multiple groups of matching XDRs and MRs in this associated time granularity can be obtained. Among them, the application record information and radio measurement report information that are not matched in this associated time granularity can be discarded.

[0091] In some possible implementation manners, for any group of matching XDR and MR, the network quality information and location information in the MR can be merged into the matching XDR to obtain a first associated record information, which can also be referred to as WQ-XDR (XDR with Wireless quality data).

[0092] In some possible implementation manners, the location information may include the grid position or number of a preset geographical location. Among them, according to the longitude and latitude of the preset geographical location, the preset geographical location can be divided into multiple grids, and the size of each grid can be, for example, 50 meters * 50 meters.

[0093] In this way, by separately processing all the matching XDRs and MRs in this associated time granularity, the first associated record information in this associated time granularity can be obtained. Furthermore, by separately performing the above processing on multiple associated time granularities, all the first associated record information within the historical time period can be obtained.

[0094] In this way, it is possible to match XDR and MR by using at least one or a combination of a time association strategy, a region association strategy, and a service type association strategy, thereby reducing the amount of matching data in multiple aspects and saving storage and computing resources; and the corresponding association strategy can improve the accuracy of matching, thereby improving the accuracy and effect of network quality difference analysis.

[0095] In some possible implementation manners, network quality difference analysis can be performed according to the first associated record information in step S13. Among them, network quality difference analysis may include identification of quality difference regions, root cause analysis of quality difference in quality difference regions, value level analysis of quality difference regions, etc.

[0096] In some possible implementation manners, the XDR includes corresponding network service quality indicators. For the identification of quality difference regions, step S13 may include: determining quality difference regions according to the network service quality indicators and the location information.

[0097] For example, multiple network service types can be distinguished according to service characteristics and content types for different applications, such as short videos, long videos, live broadcasts, page views, games, etc., and network service quality indicators for each network service can be set, such as download rate indicators, buffering time indicators, etc. For each network service quality indicator, corresponding indicator formulas can be set in combination with network quality information that can delimit wireless-side problems (for example, downlink RTT delay. If this delay is large, it is considered a problem on the wireless network side), and indicator quality difference thresholds can be set.

[0098] For example, the indicator formula for the short video playback download rate indicator is (8000 / 1024)(8000 * video playback download traffic) / (video playback download effective time * 1024); the indicator quality difference thresholds include download delay threshold T dl_tput_poor and RTT delay threshold T dl_rtt_poor . If (8000 / 1024)(8000 * video playback download traffic) / (video playback download effective time * 1024) < T dl_tput_poor and TCP downlink RTT delay > T dl_rtt_poor , it is considered that the short video playback download rate indicator has poor quality.

[0099] In some possible implementation manners, the obtained first associated record information can be processed item by item. For any first associated record information, according to the indicator formulas of each network service quality indicator, the indicator values of multiple network service quality indicators corresponding to each network service of this first associated record information can be determined respectively.

[0100] Furthermore, according to the indicator values of each network service quality indicator and the corresponding indicator quality difference thresholds, it can be determined whether each network service quality indicator has poor quality. If a certain network service quality indicator has poor quality, the indicator quality difference status field of this network service quality indicator is set to 1; conversely, if a certain network service quality indicator is normal, the indicator quality difference status field of this network service quality indicator is set to 0. In this way, the quality difference status of each network service quality indicator can be recorded into the indicator quality difference status field (referred to as S pq variable) through 1 bit.

[0101] Figure 2 This is a schematic diagram of the indicator quality difference status field provided by the embodiments of the present disclosure. As Figure 2 shown, a 64-bit indicator quality difference status field (bits 0 to 63 in Figure 2 ) can be set, corresponding to a maximum of 64 network service quality indicators (if the number of network service quality indicators is less than 64, the remaining bits are always 0). The quality difference status of each network service quality indicator of the first associated record information can be determined respectively and recorded into the indicator quality difference status field.

[0102] In some possible implementation manners, the quality difference status field (S pq value) of multiple network service quality indicators may be added to the first associated record information to obtain second associated record information. In this way, by separately processing multiple pieces of first associated record information, multiple pieces of second associated record information in a historical time period can be obtained. Through this method, each indicator of different services can be distinguished, and the quality difference status of the indicators of each service can be separately determined, thereby improving the accuracy of quality difference status determination.

[0103] In some possible implementation manners, according to multiple pieces of second associated record information and the service types of network services of a communication network, a quality difference grid where the network service has quality differences is determined from multiple grids in a preset geographical location. Among them, the second associated record information can be aggregated according to dimensions such as user information, area information, network service type information, and location information. Among them, according to multiple service types of network services, the second associated record information can be aggregated once according to a preset statistical granularity (such as an associated time granularity or the entire historical time period), and the second associated record information is divided into multiple service record information sets corresponding to each service type. The present disclosure does not limit the specific setting of the statistical granularity.

[0104] In some possible implementation manners, the service types are distinguished according to application usage service characteristics and content types, such as being divided into service types such as short video, long video, live broadcast, page browsing, and game. In the case where the historical time period is used as the statistical granularity, 5 service record information sets can be obtained; in the case where the associated time granularity is used as the statistical granularity, 5 service record information sets can be obtained for each associated time granularity. Among them, the service record information sets can be saved or displayed in the form of a service aggregation table, and the present disclosure does not limit this.

[0105] In some possible implementation manners, in the process of one aggregation, the quality difference status field S of each indicator pq can be converted into a quality difference count. For example, if there are 100 pieces of second associated record information in the service record information set of short videos, then count the number of times that the quality difference status field of the short video playback and download rate indicator is 1, for example, there are 30 times.

[0106] In some possible implementation manners, according to multiple grids in a preset geographical location, based on the location information (such as grid number) in the second associated record information, each service record information set can be secondarily aggregated, and each service record information set is divided into multiple grid record information sets corresponding to each grid. Each grid record information set includes the second associated record information of a certain service type in one grid. Among them, the service record information sets can be saved or displayed in the form of a grid aggregation table, and the present disclosure does not limit this.

[0107] In some possible implementations, during the secondary aggregation process, the quality difference ratio of the quality difference status fields of each metric corresponding to the service type of the service record information set can be calculated. For example, for the short video service, there are 100 second associated record information in a certain grid, and the number of times the quality difference status field of the short video playback and download rate metric is 1 is 30 times. Then, the quality difference ratio of the short video playback and download rate metric is 30%.

[0108] In some possible implementations, the quality difference ratio can be the quality difference ratio of a single metric or the combined quality difference ratio of multiple metrics. The combined quality difference ratio of multiple metrics can be the average of the quality difference ratios of individual metrics, or the number of times the quality difference status field is 1 is added up and the number of second associated record information is added up, and then divided. The present disclosure does not limit the specific calculation method.

[0109] In some possible implementations, if the quality difference ratio of the service type of a certain network service in a grid is greater than or equal to the regional quality difference threshold of the service type, the corresponding grid can be determined as the quality difference grid of the service type. In this way, by processing multiple grids of the preset geographical location respectively, the quality difference grids of the network service quality difference in the multiple grids of the preset geographical location can be determined.

[0110] Among them, the quality difference grids can be saved or displayed in the form of a grid quality difference table, and the present disclosure does not limit this. Moreover, the present disclosure does not limit the specific setting method and specific value of the regional quality difference threshold.

[0111] In this way, the composite metrics perceived by users are aggregated at the grid dimension, and the grids whose ratio of the number of quality difference times meets the preset conditions are screened out, so as to realize the identification of the quality difference grids for each service, improve the accuracy of identification and the precision of the determined quality difference area.

[0112] In some possible implementations, for example, the DBScan (Density-Based Spatial Clustering of Applications with Noise) method can be used to cluster the quality difference grids of the preset geographical location to obtain the quality difference areas in the preset geographical location. Among them, the reference value of the neighborhood radius parameter Epsilon for DBScan clustering can be set to 1.42, and the reference value of the data point sample number parameter Minimum Points can be set to 2. It should be understood that those skilled in the art can set the specific clustering method and parameters according to the actual situation, and the present disclosure does not limit this.

[0113] Figure 3 This is a schematic diagram of the quality difference area clustering provided by the embodiments of the present disclosure. AsFigure 3 As shown in the figure, the right figure shows multiple poor-quality grids that have been recognized. After clustering, two poor-quality regions in the left figure are obtained, which are assigned values of 1 and 2, and the isolated poor-quality grids are uniformly identified as -1. Among them, each poor-quality region includes at least two adjacent poor-quality grids.

[0114] In this way, adjacent poor-quality grids can be associated to obtain a poor-quality region that is more in line with the actual situation, improving the accuracy of poor-quality region division.

[0115] In some possible implementation manners, the network poor-quality analysis further includes root cause analysis of the poor-quality region. Step S13 may further include: determining the root cause of the poor-quality region according to the network quality information.

[0116] In the embodiments of the present disclosure, model training may be first performed according to historical data to identify wireless factors that can affect service quality and poor-quality thresholds that cause a decline in the wireless service experience, and classification rules for the root cause of the poor-quality are determined; then, according to the classification rules for the root cause of the poor-quality, the category of the root cause of the poor-quality region is determined.

[0117] In some possible implementation manners, according to the network poor-quality analysis method of the embodiments of the present disclosure, it further includes:

[0118] Determining a mapping list of the network quality information in the MR and the network service quality indicators in the XDR according to the first association record information in the historical time period;

[0119] Through a dichotomy based on the statistic F, the mapping list of the network quality information and the network service quality indicators is traversed by dichotomy to determine the first index value of the network quality information for the network service quality indicators;

[0120] Clustering the first index values of the network quality information for multiple network service quality indicators to obtain the poor-quality threshold of the network quality information;

[0121] Determining a regression model of the network quality information in the MR for the network service quality indicators in the XDR according to the first association record information of the poor-quality region in the historical time period;

[0122] Determining the confidence rule of the network quality information according to the regression model and the poor-quality threshold of the network quality information.

[0123] For example, the first association record information in the historical time period can be obtained, the network quality information (such as RTT delay, 5G signal strength, etc.) in the MR and the network service quality indicators (such as video download rate, etc.) in the XDR in the first association record information can be extracted, and a mapping list of the network quality information and the network service quality indicators is established.

[0124] In some possible implementation manners, it is possible to determine whether the p-value is less than the significance level of 0.05 based on one-way analysis of variance, construct a set of wireless root causes leading to poor service quality, where the set of wireless root causes includes multiple network quality information, such as poor downlink coverage; and adopt a dichotomy based on the statistic F to automatically learn the quality difference threshold at which the wireless quality difference affects the service perception. According to the occurrence probability of each factor in the root cause set and the significance level, the confidence level of the root cause is evaluated.

[0125] Figure 4 This is a schematic diagram of the learning of the quality difference threshold provided by the embodiments of the present disclosure. As Figure 4 shown, the learning of the quality difference threshold of the network quality information may include the following steps:

[0126] According to a preset time dimension (such as 1 day, 1 hour, etc.), extract the network quality information and network service quality indicators from the first associated record information, construct a mapping list of the network quality information and network service quality indicators; sort the network quality information from good to bad to obtain a sorted mapping list.

[0127] Through the dichotomy based on the statistic F, perform a binary split on the sorted mapping list, and calculate the statistic F for the two sub-lists obtained by the split; put the statistic F into an independent statistic list. In this way, perform a binary split traversal on the sorted mapping list, that is, split each position of the mapping list in turn, calculate the statistic F of the sub-list, and put it into the statistic list.

[0128] After the binary split traversal ends, search for the index of the binary split position with the largest statistic F value in the statistic list; determine the first index value of the corresponding network quality information in the mapping list based on the position index as the quality difference reference value.

[0129] In this way, by performing the above processing on the historical data of multiple time dimensions (for example, taking the historical data of each day or each hour as the historical data of one time dimension), multiple first index values can be obtained as the quality difference reference values.

[0130] Perform clustering processing on multiple first index values, such as using K-means clustering, cluster multiple first index values into a cluster, and use the center point of the clustered cluster as the quality difference threshold of the network quality information in the quality difference root cause classification rule.

[0131] In some possible implementation manners, according to the first associated record information of the quality difference area in the historical time period, based on the set of wireless root causes leading to poor service quality, statistically analyze the relationship between the network quality information leading to wireless quality difference and the network service quality indicators leading to poor service quality, create a quadratic polynomial regression model y = f(x) of the network quality information in MR for the network service quality indicators in XDR, and then take the derivative of the quadratic polynomial to obtain the extreme point (xm , y m ); Based on the quality difference threshold of network quality information, the threshold point (x t , y t ) is obtained through a quadratic polynomial regression model. Then, the formula corresponding to the confidence rule of network quality information is as follows:

[0132] R confidence = a + (100 - a) * (f(x) - y t ) / (y m - y t ) (2)

[0133] In formula (2), R confidence represents the confidence level, a represents a preset parameter, which is an empirical value, such as 60%, and x is the value of network quality information.

[0134] Figure 5 is a schematic diagram of the relationship between network quality information and network service quality indicators provided by an embodiment of the present disclosure. As Figure 5 shown, the network quality information that causes service quality difference is, for example, downlink weak coverage, and the network service quality indicator is, for example, download rate; Figure 5 The solid line in represents the proportion of download rate quality difference, and the dotted line represents the proportion of download rate quality difference obtained through quadratic polynomial regression. The vertical axis represents the ratio of the quality difference proportion, with a range of 0 to 0.08; the horizontal axis is the indicator value of downlink weak coverage, with a range of 0 to 1. It can be seen that after being processed by the quadratic polynomial regression model, the relationship between network quality information and network service quality indicators is smoother.

[0135] In some possible implementation manners, the root causes of quality difference in the quality difference area may include any one of the following: network coverage quality difference, network interference quality difference, network capacity quality difference, network performance quality difference, and network fault quality difference. The present disclosure does not limit this.

[0136] In the example, for the network coverage quality difference and network interference quality difference among the root causes of quality difference, according to the first associated record information, the relationship between each network quality information and network service quality indicator can be converted into discrete features in the form of bucketing, and in the form of a histogram, as the input data for regression model training. For the network capacity quality difference and network performance quality difference among the root causes of quality difference, the network quality information and network service quality indicator in the first associated record information can be divided into data in the dimension of radio cells, and regression model training is performed based on the data in the dimension of radio cells. After training, according to the training results of the regression model, a rule library for network quality difference root cause location is generated, and the rule library includes multiple quality difference root cause classification rules. It should be understood that those skilled in the art can set the root cause model and quality difference root cause according to the actual situation, and the present disclosure does not limit this.

[0137] In some possible implementation manners, in step S13, determining the root cause of the poor quality area according to the network quality information includes:

[0138] Determining the first network quality information that causes the poor quality and the confidence level of the first network quality information according to the network quality information of the poor quality area, the poor quality threshold of each network quality information, and the confidence rule;

[0139] Determining the root cause of the poor quality area according to the confidence level of the first network quality information and the root cause of the poor quality corresponding to the first network quality information.

[0140] For example, each poor quality area in the preset geographical location can be processed separately. For any poor quality area, the corresponding radio cell of the poor quality area can be determined first; and it can be determined whether the radio alarm data in the network element network management performance data matches the radio cell corresponding to the poor quality area; if it matches, the root cause of the poor quality of the poor quality area can be directly determined as the network failure poor quality; on the contrary, if it does not match, other root cause categories of the poor quality can be further judged.

[0141] In some possible implementation manners, according to the radio cell corresponding to the poor quality area, a service aggregation table aggregated once can be associated, that is, a set of service record information of the preset geographical location in the historical time period, and then the first associated record information corresponding to the poor quality area can be determined; according to the network quality information in the first associated record information and the poor quality threshold of each network quality information, at least one first network quality information (such as weak coverage) that causes the service poor quality of the poor quality area can be determined; furthermore, according to the confidence rule of the network quality information, the confidence level of each first network quality information can be determined; according to the root cause of the poor quality corresponding to the network quality information with the highest confidence level, the root cause of the poor quality area can be determined. The root cause of the poor quality includes any one of network coverage poor quality, network interference poor quality, network capacity poor quality, and network performance poor quality. In this way, by performing the above processing on each poor quality area of the preset geographical location, the root cause of each poor quality area can be obtained.

[0142] In this way, the accuracy and objectivity of the determination of the root cause of the poor quality area can be improved, and the subjectivity problem existing in the case of manual determination can be avoided.

[0143] After the identification of the poor quality area is completed, the value level of the poor quality area can also be analyzed to provide a basis for the operation and maintenance personnel, so that the operation and maintenance personnel can intuitively judge the influence of the poor quality area and the optimization value of the poor quality area. Among them, the value of the poor quality area can be comprehensively evaluated based on several factors such as the coverage range of the poor quality area, the number of affected users, the traffic volume, the degree of service poor quality, and the improvement cost. The higher the value, the greater the benefit of optimizing the poor quality area.

[0144] In some possible implementation manners, the coverage range of a poor quality area can be evaluated by the number of poor quality grids corresponding to the poor quality area. The more the number of poor quality grids, the higher the value of the poor quality area. The number of users affected by the poor quality area can be evaluated by the total number of users in the poor quality area, that is, the number of terminals corresponding to the second association record information of the poor quality of each service type. The more the total number of users, the higher the value of the poor quality area. The traffic volume of the poor quality area can be evaluated by the number of service occurrences, that is, the number of records of the second association record information of each service type. The more the number of service occurrences, the higher the value of the poor quality area. The degree of service quality degradation in the poor quality area can be evaluated by the service quality degradation ratio, that is, the service quality degradation ratio of the second association record information of the poor quality of each service type. The higher the service quality degradation ratio, the higher the value of the poor quality area. The improvement cost of the poor quality area can be evaluated by the ratio of the number of poor quality grids in the poor quality area to the number of poor quality radio cells corresponding to the poor quality area. The larger this ratio, the more the poor quality is concentrated in one or several cells, and the higher the value of the poor quality area.

[0145] In some possible implementation manners, the method according to the embodiments of the present disclosure further includes:

[0146] For any poor quality area of the preset geographical location, according to the second association record information of the poor quality grid corresponding to the poor quality area, determine the area parameters of the poor quality area; the area parameters include at least one of the following: the number of poor quality grids corresponding to the poor quality area, the number of terminals corresponding to the second association record information of the poor quality of each service type, the number of records of the second association record information of each service type, the service quality degradation ratio of the second association record information of the poor quality of each service type, the ratio of the number of poor quality grids to the number of poor quality radio cells corresponding to the poor quality area;

[0147] According to each area parameter of the poor quality area, as well as the interval threshold and parameter weight of each area parameter, determine the value score of the poor quality area;

[0148] According to the value score of the poor quality area, determine the value level of the poor quality area.

[0149] For example, each poor quality area of the preset geographical location can be processed separately. For any poor quality area, first determine the area parameters of the poor quality area. The area parameters include at least one of the following: the number of poor quality grids corresponding to the poor quality area, the number of terminals corresponding to the second association record information of the poor quality of each service type, the number of records of the second association record information of each service type, the service quality degradation ratio of the second association record information of the poor quality of each service type, the ratio of the number of poor quality grids corresponding to the poor quality area to the number of poor quality radio cells corresponding to the poor quality area.

[0150] In the example, the poor-quality area is determined by clustering poor-quality grids, and the number of poor-quality grids corresponding to the poor-quality area can be directly determined.

[0151] In the example, according to the second association record information of the poor-quality grids corresponding to the poor-quality area, the users corresponding to the second association record information of the poor-quality of each service type can be determined, that is, the terminals corresponding to the second association record information; after de-duplicating the terminals corresponding to the second association record information, the corresponding number of terminals can be determined.

[0152] In the example, according to the second association record information of the poor-quality grids corresponding to the poor-quality area, the record quantity of the second association record information of each service type can be determined, that is, the number of service times of each service type.

[0153] In the example, according to the second association record information of the poor-quality grids corresponding to the poor-quality area, the poor-quality proportion of the second association record information of the poor-quality of each service type can be determined. Among them, the poor-quality proportion of each grid has been determined respectively in the previous steps, and the poor-quality proportion of this poor-quality area can be calculated according to this poor-quality proportion; it is also possible to sum the quantities of the second association record information of the poor-quality of each service type of each grid in this poor-quality area, and compare it with the total quantity of the second association record information of each grid in this poor-quality area, and use the obtained ratio as the poor-quality proportion of this poor-quality area. The present disclosure does not limit this.

[0154] In the example, according to the positions of the poor-quality grids corresponding to the poor-quality area and the positions of the radio cells of the base station, the number of radio cells corresponding to the poor-quality area can be determined; furthermore, the number of poor-quality grids corresponding to the poor-quality area is compared with the number of radio cells, and the ratio of the number of poor-quality grids corresponding to the poor-quality area to the number of poor-quality radio cells corresponding to the poor-quality area is obtained.

[0155] In some possible implementation manners, the interval threshold and parameter weight of each area parameter can also be determined. The interval threshold can be determined according to the actual data distribution of the area parameter, and the parameter weight can be determined by verifying according to practical experience. The present disclosure does not limit the specific determination method and value of the interval threshold and parameter weight.

[0156] In the example, let the interval threshold of each area parameter be C min and C max , C min and C max be the minimum value and the maximum value of the interval threshold respectively, then the score S of each area parameter is as follows:

[0157]

[0158] In formula (3), C represents the specific value of the area parameter, C minand C max are respectively the minimum and maximum values of the interval thresholds of the regional parameters.

[0159] In the example, when the regional parameters are: the number of quality difference grids corresponding to the quality difference region, the number of terminals corresponding to the second associated record information of the quality difference of each service type, the number of records of the second associated record information of each service type, the proportion of the quality difference of the second associated record information of each service type, and the ratio of the number of quality difference grids corresponding to the quality difference region to the number of quality difference radio cells corresponding to the quality difference region, these 5 parameters, the parameter weights can be W = [0.1, 0.3, 0.2, 0.25, 0.15].

[0160] In some possible implementation manners, the score S of each regional parameter can be determined according to the interval threshold respectively, and then the weighted sum of the scores can be calculated according to the parameter weights to obtain the value score V of the quality difference region 价值 :

[0161]

[0162] In formula (4), k represents the number of the regional parameter, n is the number of regional parameters, W k represents the parameter weight of the k-th regional parameter, and S k represents the score of the k-th regional parameter.

[0163] In some possible implementation manners, after obtaining the value score of the quality difference region, the value level of the quality difference region can be determined according to. For example, according to the value of the value score, in the four intervals of [90, 100], [80, 90), [60, 80), [0, 60), the value level of the quality difference region is divided into four levels: excellent, good, medium, and poor. In this way, according to the value score of the quality difference region, the value level of the quality difference region can be determined, for example, as good. It should be understood that those skilled in the art can set the calculation method of the value score of the quality difference region and the classification method of the value level according to the actual situation, and the present disclosure does not limit this.

[0164] In this way, an intuitive value score and value level can be given to the quality difference region, so that the operation and maintenance personnel can intuitively judge the influence of the quality difference region and the optimization value of the quality difference region, thereby improving the efficiency of operation and maintenance.

[0165] The implementation process of the method according to the embodiments of the present disclosure will be described below through embodiments.

[0166] In the example, the service type to be analyzed is the short video service of smartphone mobile Internet access. The implementation steps of this embodiment are as follows:

[0167] Step 1: Associate the application record information and the radio measurement report information.

[0168] In the example, first, the parameters in the association policy for application record information (XDR) and radio measurement report information (MR) are configured as shown in Table 2 below.

[0169] Table 2

[0170] Parameter Name Abbreviation Value Associated Time Granularity <![CDATA[G r > 1 hour XDR Delay Threshold <![CDATA[D xdr > 30 MR Delay Threshold <![CDATA[D mr > 15 Key Services <![CDATA[P app > (5,65) Preset Geographical Location <![CDATA[P zone > (0,531) Busy Hour Time Range <![CDATA[T busy > [7,23] Associated Time Window <![CDATA[W a > 60 seconds

[0171] In the example, the radio cell configuration is associated through the regional range (0, 531) of the preset geographical location to obtain the set P of radio cells cell ={c1, c2, c3,...}, and the application record information reported from the external system is received. The application record information in this embodiment is the detailed list of video playback facts.

[0172] In the example, based on the association policy, according to P app , P cell and T busy the application record information is filtered out, saved to the partition folder, and the moment T when the application record information is received last in the current time granularity (i.e., the association time granularity) is updated xdr , and it is regularly detected whether the data in the current time granularity is complete. For example, when the current time is T current , when T current -T xdr >D xdr then the application record information in the current time granularity is turned over.

[0173] In the example, the radio measurement report information (MR detailed list) reported from the external system is received. Based on the association policy, according to P cell and T busy the radio measurement report information is filtered out, saved to the partition folder, and the moment T when the radio measurement report information is received last in the current time granularity is updated mr , and it is regularly detected whether the data in the current time granularity is complete. For example, when the current time is T current , when T current -T mr >D mr then the radio measurement report information in the current time granularity is turned over.

[0174] In the example, after the application record information and the radio measurement report information in the current time granularity are both turned over, a matching operation is performed. The matching process uses the user identifier and the radio cell identifier as keywords and needs to meet the following conditions:

[0175] |XDR.T proc +XDR.duration / 2 - MR.T proc |<=W a (1)

[0176] In formula (1), XDR.T proc represents the first service occurrence time in the application record information; XDR.duration represents the service duration; MR.T proc represents the second service occurrence time in the radio measurement report information; W a represents the associated time window.

[0177] In the example, after preliminary matching, one application record information may match multiple alternative radio measurement report information. First, according to the positioning type of the radio measurement report information (such as GPS positioning, server positioning (such as triangulation), etc.), the radio measurement report information with higher positioning type accuracy (such as GPS positioning) is preferentially selected; if the positioning types are the same, the radio measurement report information with a smaller associated time difference is preferentially selected, and finally, one radio measurement report information that matches the application record information can be determined, and a set of matching XDR and MR can be obtained.

[0178] In the example, the network quality information and location information in the radio measurement report information can be merged into the matching XDR to obtain a first associated record information, called WQ-XDR.

[0179] In this way, by processing all the matching XDR and MR in this associated time granularity respectively, the first associated record information of this associated time granularity can be obtained. Furthermore, by performing the above processing on multiple associated time granularities in the historical time period respectively, the first associated record information of the historical time period can be obtained.

[0180] Step 2: Identification of index quality degradation.

[0181] In the example, first, based on the first associated record information, it is identified whether the network service quality index is of poor quality. Taking the short video playback and download rate index as an example, the index formula is (8000 / 1024)(8000 * video playback and download traffic) / (video playback and download effective time * 1024); the index quality degradation thresholds include the download delay threshold T dl_tput_poor and the RTT delay threshold T dl_rtt_poor . If (8000 / 1024)(8000 * video playback and download traffic) / (video playback and download effective time * 1024) < T dl_tput_poor and the TCP downlink RTT delay > T dl_rtt_poor , it is considered that the short video playback and download rate index is of poor quality.

[0182] In this case, the index quality degradation status field, that is, the S pq variable, sets the bit corresponding to the short video playback and download rate index to 1. For example, S pq= 0x0000000000000001 (the last digit is the short video playback and download rate indicator). In this way, the network service quality indicators of each service are processed separately to determine the values of each bit of the S pq variable; then the S pq variable is filled back into the first associated record information (WQ-XDR) to obtain the second associated record information.

[0183] In this way, by performing the above processing on the first associated record information of multiple associated time granularities in the historical time period, the second associated record information in the historical time period can be obtained.

[0184] Step 3: Primary aggregation.

[0185] In the example, based on dimensions such as user information, region information, service information, and location information, the second associated record information is primarily aggregated according to a preset statistical granularity (such as 1 day). The aggregation table distinguishes major service categories, and the major service categories are distinguished according to the service characteristics and content types of the application. During the aggregation process, the S pq is converted into the statistical count of the quality difference times of the corresponding indicator. The situation of the video playback and download rate service on a certain day is shown in the primary aggregation table of Table 3 below.

[0186] Table 3

[0187]

[0188] Step 4: Secondary aggregation and determination of the quality difference area.

[0189] In the example, based on the primary aggregation table, secondary aggregation is performed according to the grid dimension to form a grid aggregation table. The quality difference ratio of all network service quality indicators is calculated, and the comprehensive quality difference ratio is calculated according to the ratio of each indicator. According to the grid quality difference threshold, the quality difference grids are screened from the grid aggregation table to form a grid quality difference table.

[0190] In the example, AI clustering can be performed on the grid quality difference table based on DBScan to identify adjacent quality difference grids as quality difference areas and assign a unique numerical number to them. Isolated grids are uniformly numbered -1. The reference value of the neighborhood radius parameter Epsilon for DBScan clustering is 1.42, and the reference value of the data point sample number parameter Minimum Points is 2. The generated table of quality difference areas (grid black dot table) is shown in Table 4:

[0191] Table 4

[0192] Start Time 202X-8-2X 202X-8-2X Statistical Granularity 86400 86400 City City 1 City 1 Grid (5041,-27,-9) (5041,-27,-8) Poor Quality Area Identifier 1000 1000 Traffic Volume 75 50 Number of Video Playback and Download Rate Services 71 48 Number of Poor Quality Services for Video Playback and Download Rate 15 15

[0193] Step 5: Root cause analysis of the quality difference area.

[0194] In the example, based on the association between the grid black point table and the primary convergence table, according to the preset classification rules for quality degradation root causes, the root causes of quality degradation in the quality degradation area are found. The black point root cause table is formed as shown in Table 5 below:

[0195] Table 5

[0196] Start Time 202X-8-2X Statistical Granularity 86400 City City 1 Poor Quality Area Identifier 1000 Cell Identifier 6802432705FE001 Root Cause Identifier 201 <Weak coverage> Confidence Level 87％

[0197] Step 6: Value assessment of the quality degradation area.

[0198] In the example, based on the black point root cause table, a black point library is generated through comprehensive evaluation. The interval thresholds of the preset scores for the black point coverage range, the number of affected users, the traffic volume, the degree of service quality degradation, and the improvement cost are [1, 3], [1, 5], [10, 100], [2, 20], [0.5, 3] respectively. Combining the weights W of each factor, according to formulas (3) and (4), the value score and value level of the quality degradation area are calculated, as shown in the following table:

[0199] Table 6

[0200] Start Time 202X-8-2X Statistical Granularity 86400 City City 1 Poor Quality Area Identifier 1000 Number of Grids 2 Number of Affected Users 2 Traffic Volume 125 Degree of Service Poor Quality 27.73％ Improvement Cost 2 Value Score 66.5 Value Level Medium

[0201] In this way, by performing the above processing on each service type in each quality degradation area of the preset geographical location respectively, the root causes of quality degradation and the value levels of the quality degradation areas can be obtained, completing the entire processing process.

[0202] According to the network quality degradation analysis method of the present disclosure, a flexible association and screening strategy is set to reduce the amount of data to be processed, which can save storage and computing resources; through intelligent analysis to identify quality degradation areas, perform black point root cause analysis and black point value assessment, etc., it can enable the identification and analysis of network black points to be truly applied to production, improve operation and maintenance efficiency, shorten the fault recovery cycle, and enhance user satisfaction.

[0203] According to the embodiments of the present disclosure, based on flexible association strategies such as time association strategy, region association strategy, and service type association strategy, on-demand association of XDR and MR can be achieved, saving a large amount of storage and computing resources. At the same time, the screening strategy is dynamically adjusted. In the time dimension, polling scheduling is performed to exchange time for space, which not only saves resources but also meets the requirements of network quality degradation analysis for all scenarios.

[0204] According to an embodiment of the present disclosure, a preset geographical location can be divided into multiple grids. Based on the grid dimension, clustering is performed on the perception composite index, and grids with the proportion of the number of times of poor quality of the index meeting the preset conditions are screened out. The grids are intelligently mined through the DBScan AI algorithm, and the grids with poor quality are merged to identify the areas with poor quality. Each area with poor quality is uniquely identified, and clustering is performed on indicators such as the number of users, traffic volume, proportion of the number of times of poor service quality, and proportion of poor radio quality based on the identifier of the area with poor quality, automatically completing the network quality analysis and improving the operation and maintenance efficiency.

[0205] According to an embodiment of the present disclosure, based on the one-way analysis of variance, the root cause set of service quality problems can be determined with the significance level of the impact of radio quality on service perception. The quality threshold affecting service perception is automatically learned based on the dichotomy of the statistic F. Quadratic polynomial regression is performed according to the occurrence probability of each factor in the root cause set to realize the confidence evaluation of multiple root causes. Rule learning is carried out based on big data, thereby improving the accuracy and objectivity of the determination of the root cause of the quality problem in the area with poor quality and avoiding the subjectivity problem existing in the case of manual determination.

[0206] According to an embodiment of the present disclosure, based on the radio quality index, the proportion of radio poor quality MR can be statistically calculated. Based on objective decision rules, the radio root cause of service quality problems can be deduced more accurately, effectively supporting the optimization of the area with poor quality. Moreover, based on factors such as the black spot coverage range, the number of users, traffic volume, service quality degradation degree, and improvement cost, the value of the area with poor quality can be evaluated based on weights and linear scoring, and an intuitive value level of the area with poor quality is given, enabling operation and maintenance personnel to intuitively judge the impact of the area with poor quality and the value of network optimization.

[0207] The network quality analysis method according to an embodiment of the present disclosure can perform intelligent analysis through the signaling data of the mobile network, discover the areas with poor quality in the mobile network coverage, effectively solve the shortcomings of the prior art, enable the areas with poor quality to be truly applied in production, improve the operation and maintenance efficiency, shorten the fault recovery cycle, and improve user satisfaction.

[0208] Figure 6 It is a block diagram of a network quality analysis device provided by an embodiment of the present disclosure. Referring to Figure 6 , an embodiment of the present disclosure also provides a network quality analysis device, which includes:

[0209] An information acquisition module 61, configured to acquire application record information XDR and radio measurement report information MR of a communication network to be analyzed;

[0210] An information matching module 62, configured to match the XDR and the MR, and merge the network quality information and location information in the MR into the XDR matched with the MR to obtain the first associated record information;

[0211] The poor quality analysis module 63 is used to analyze the network poor quality according to the first associated record information.

[0212] In some possible implementation manners, the information matching module 62 includes matching the XDR and the MR according to at least one of the following association policies or a combination thereof:

[0213] The time association policy is used to match the XDR and the MR based on a preset time policy;

[0214] The area association policy is used to match the XDR and the MR based on a preset geographical location;

[0215] The service type association policy is used to match the XDR and the MR based on a preset service type.

[0216] In some possible implementation manners, the time policy includes at least one of the following:

[0217] The associated time granularity is used to represent the time interval corresponding to the XDR and the MR participating in the matching;

[0218] The XDR delay threshold is used to represent the maximum delay of the reporting time of the XDR relative to the time interval;

[0219] The MR delay threshold is used to represent the maximum delay of the reporting time of the MR relative to the time interval;

[0220] The busy time range is used to represent the busy time range of the communication network in a day;

[0221] The associated time window is used to represent the time difference range between the matching XDR and MR.

[0222] In some possible implementation manners, the poor quality analysis module 63 is used to: determine a poor quality area according to the network service quality index and the location information.

[0223] In some possible implementation manners, the poor quality analysis module 63 is used to: determine the root cause of the poor quality of the poor quality area according to the network quality information.

[0224] In some possible implementation manners, the poor quality analysis module 63 is used to: determine the first network quality information causing the poor quality and the confidence level of the first network quality information according to the network quality information of the poor quality area, the poor quality threshold of each network quality information, and the confidence rule; determine the root cause of the poor quality of the poor quality area according to the confidence level of the first network quality information and the root cause of the poor quality corresponding to the first network quality information.

[0225] In some possible implementations, the apparatus further includes:

[0226] A mapping list determination module, configured to determine a mapping list between network quality information in the MR and network service quality metrics in the XDR according to first association record information in a historical time period;

[0227] An index value determination module, configured to perform binary segmentation traversal on the mapping list between the network quality information and the network service quality metrics by means of a dichotomy based on statistic F, and determine a first index value of the network quality information for the network service quality metrics;

[0228] A quality difference threshold determination module, configured to cluster first index values of the network quality information for multiple network service quality metrics to obtain a quality difference threshold of the network quality information;

[0229] A model determination module, configured to determine a regression model of network quality information in the MR for network service quality metrics in the XDR according to first association record information of a quality difference region in the historical time period;

[0230] A confidence determination module, configured to determine a confidence rule of the network quality information according to the regression model and the quality difference threshold of the network quality information.

[0231] Each module in the foregoing apparatus may be implemented in whole or in part by software, hardware, and a combination thereof. Each of the foregoing modules may be embedded in a processor in a computer device in a hardware form or be independent of the processor, or may be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the foregoing modules.

[0232] Figure 7 This is a block diagram of a composition of an electronic device provided by an embodiment of the present disclosure. An embodiment of the present disclosure further provides an electronic device, as Figure 7 shown, including:

[0233] At least one processor 701;

[0234] A memory 702, on which at least one program is stored. When the at least one program is executed by the at least one processor, the at least one processor implements the network quality difference analysis method provided in the foregoing embodiments;

[0235] At least one I / O interface 703, connected between the processor and the memory, and configured to implement information interaction between the processor and the memory.

[0236] Among them, the processor 701 is a device with data processing capabilities, including but not limited to a central processing unit (CPU), etc.; the memory 702 is a device with data storage capabilities, including but not limited to a random access memory (RAM, more specifically such as SDRAM, DDR, etc.), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), and a flash memory (FLASH); the I / O interface (read / write interface) 703 is connected between the processor 701 and the memory 702 and can implement information interaction between the processor 701 and the memory 702, including but not limited to a data bus (Bus), etc.

[0237] In some embodiments, the processor 701, the memory 702, and the I / O interface 703 are interconnected via a bus and are further connected to other components of the computing device.

[0238] The embodiments of the present disclosure also provide a computer-readable medium, on which a computer program is stored, where when the computer program is executed, it implements the network quality degradation analysis method provided in the foregoing embodiments.

[0239] Those of ordinary skill in the art can understand that all or some of the steps, systems, and functional modules / units in the devices disclosed above can be implemented as software, firmware, hardware, and their appropriate combinations.

[0240] In the hardware implementation, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component can have multiple functions, or a function or step can be executed by several physical components in cooperation.

[0241] Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit (CPU), a digital signal processor or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-temporary medium) and a communication medium (or temporary medium). As known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory (FLASH) or other disk storage; read-only compact disk (CD-ROM), digital versatile disk (DVD) or other optical disk storage; magnetic cassettes, magnetic tapes, disk storage or other magnetic storage; any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0242] The present disclosure has disclosed example embodiments, and although specific terms are employed, they are used and should be interpreted only in a general illustrative sense and not for limiting purposes. In some instances, it will be apparent to those skilled in the art that, unless otherwise expressly stated, features, characteristics, and / or elements described in conjunction with a particular embodiment may be used alone or in combination with features, characteristics, and / or elements described in conjunction with other embodiments. Therefore, those skilled in the art will appreciate that various changes in form and detail may be made without departing from the scope of the present disclosure as set forth in the appended claims.

Claims

1. A method for analyzing poor network quality, which includes: Obtaining the application record information XDR and wireless measurement report information MR of the communication network to be analyzed; Matching the XDR and the MR, and merging the network quality information and location information in the MR into the XDR that matches the MR to obtain the first associated record information; Analyzing the poor network quality according to the first associated record information.

2. The method according to claim 1, wherein, The matching of the XDR and the MR includes matching the XDR and the MR according to at least one or a combination of the following association strategies: Time association strategy, matching the XDR and the MR based on a preset time strategy; Region association strategy, matching the XDR and the MR based on a preset geographical location; Service type association strategy, matching the XDR and the MR based on a preset service type.

3. The method according to claim 2, wherein, The time strategy includes at least one of the following: Association time granularity, used to characterize the time interval corresponding to the XDR and MR participating in the matching; XDR delay threshold, used to characterize the maximum delay of the reporting time of the XDR relative to the time interval; MR delay threshold, used to characterize the maximum delay of the reporting time of the MR relative to the time interval; Busy time range, used to characterize the busy time range of the communication network in a day; Association time window, used to characterize the time difference range between the matching XDR and MR.

4. The method according to claim 1, wherein, The XDR includes corresponding network service quality indicators; the analyzing the poor network quality according to the first associated record information includes: Determining the poor quality area according to the network service quality indicators and the location information.

5. The method according to claim 4, wherein, The analyzing the poor network quality according to the first associated record information includes: Determining the root cause of the poor quality in the poor quality area according to the network quality information.

6. The method according to claim 5, wherein, The determining the root cause of the poor quality in the poor quality area according to the network quality information includes: Determining the first network quality information causing the poor quality and the confidence level of the first network quality information according to the network quality information in the poor quality area, the poor quality threshold of each network quality information, and the confidence rule; Determining the root cause of the poor quality in the poor quality area according to the confidence level of the first network quality information and the root cause of the poor quality corresponding to the first network quality information.

7. The method according to claim 6, wherein, The method further includes: Determining a mapping list of the network quality information in the MR and the network service quality indicators in the XDR according to the first associated record information in the historical time period; Through the dichotomy based on the statistic F, performing a binary split traversal on the mapping list of the network quality information and the network service quality indicators to determine the first index value of the network quality information for the network service quality indicators. Cluster the first index values of the network quality information for multiple network service quality indicators to obtain the quality difference threshold of the network quality information; Determine the regression model of the network quality information in the MR for the network service quality indicators in the XDR according to the first associated record information of the quality difference area in the historical time period; Determine the confidence rule of the network quality information according to the regression model and the quality difference threshold of the network quality information.

8. A network quality difference analysis device, which includes: An information acquisition module, configured to acquire application record information XDR and radio measurement report information MR of a communication network to be analyzed; An information matching module, configured to match the XDR and the MR, and merge the network quality information and location information in the MR into the XDR matched with the MR to obtain first associated record information; A quality difference analysis module, configured to analyze network quality differences according to the first associated record information.

9. An electronic device, which includes a memory and a processor; the memory stores a computer program executable by the processor, and when the computer program is executed by the processor, it implements the network quality difference analysis method according to any one of claims 1 to 7.

10. A computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the network quality difference analysis method according to any one of claims 1 to 7.

Images