Wireless index detection method, storage medium and electronic device

Abstract

The embodiments of the present invention provide a wireless indicator detection method, a storage medium and an electronic device, which selects the detection model with the highest priority of model detection accuracy as the first detection model for detection according to the detection request; and trains the detection model with a higher priority of model detection accuracy than the first detection model according to the communication data of the wireless indicator to be detected, so as to detect the next detection request of the same type. This can solve the problem of large training amount and low accuracy of the detection model in the wireless indicator detection in the related technology, and achieve the effect of reducing the training amount of the detection model and improving the accuracy of the wireless indicator detection.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of communications, and in particular to a wireless indicator detection method, a storage medium and an electronic device. Background Art

[0002] In recent years, as the operation and maintenance system continues to develop in the direction of intelligence, artificial intelligence (AI) algorithms are increasingly embedded in operators' production systems. Wireless operation and maintenance scenarios include indicator prediction, anomaly warning, root cause analysis, anomaly detection, etc. For wireless operation and maintenance detection algorithms, the current common practice is to first train a relevant wireless operation and maintenance model, predict the key performance indicators (KPIs) for the next N days, and then use supervised or unsupervised methods to test the predicted values, perform indicator prediction or anomaly warning based on the test results, and also perform root cause analysis or anomaly location based on the relevant wireless operation and maintenance model combined with actual data.

[0003] The bottleneck of the above wireless operation and maintenance methods is usually that the model needs to be trained in advance so that it can respond in time during reasoning and meet the real-time requirements of the production system. However, whether it is to train a general model in advance or to train a separate model for each indicator or object, there are problems: there are hundreds of indicators in the field of wireless operation and maintenance, and the objects under different dimensional combinations are increased by orders of magnitude. It undoubtedly requires a huge workload to train and maintain a model for each indicator or even object in advance. When new objects appear in the system, the model needs to be trained before use, and the scalability is very poor. Although training only a general model can solve the problem of needing to train a large number of models, its accuracy is greatly reduced. The more indicators and objects there are, the worse the model performance. If a large model with a large number of parameters is used to improve the accuracy, it will increase the difficulty of model design and the consumption of computing resources. At the same time, it will increase the difficulty of reasoning, making it difficult for the system to return the reasoning results in time.

[0004] In summary, the existing wireless operation and maintenance detection models have the problems of large training volume and low accuracy. Summary of the invention

[0005] The embodiments of the present invention provide a wireless indicator detection method, a storage medium and an electronic device to at least solve the problem that the detection model in the wireless indicator detection in the related art has a large amount of training and low accuracy.

[0006] According to one embodiment of the present invention, a wireless indicator detection method is provided, including: selecting an existing detection model with the highest model detection accuracy priority as a first detection model for detection according to a detection request; and training a detection model with a model detection accuracy priority higher than that of the first detection model according to communication data of the wireless indicator to be detected, so as to detect the next detection request of the same type.

[0007] According to yet another embodiment of the present invention, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the steps of any one of the above method embodiments when run.

[0008] According to yet another embodiment of the present invention, there is provided an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0009] Through the present invention, a wireless indicator detection method is provided, which selects the detection model with the highest priority of model detection accuracy as the first detection model for detection according to the detection request; and trains the detection model with a higher priority of model detection accuracy than the first detection model according to the communication data of the wireless indicator to be detected, so as to detect the next detection request of the same type. The problem of large training amount and low accuracy of the detection model in the wireless indicator detection in the related technology can be solved, so as to achieve the effect of reducing the training amount of the detection model and improving the accuracy of the wireless indicator detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 It is a hardware structure block diagram of a computer terminal of a wireless index detection method according to an embodiment of the present invention;

[0011] Figure 2 is a schematic diagram of a wireless indicator detection operation network architecture according to an embodiment of the present invention;

[0012] Figure 3 is a flow chart of a wireless indicator detection method according to an embodiment of the present invention;

[0013] Figure 4 is a structural block diagram of a wireless indicator detection device according to an embodiment of the present invention;

[0014] Figure 5 is a schematic diagram of a model reasoning process according to a scenario embodiment of the present invention;

[0015] Figure 6 is a schematic diagram of the reasoning process of the detection algorithm according to a scenario embodiment of the present invention;

[0016] Figure 7 is a schematic diagram of a model training process according to a scenario embodiment of the present invention;

[0017] Figure 8 It is a schematic diagram of the detection algorithm training process according to a scenario embodiment of the present invention. DETAILED DESCRIPTION

[0018] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings and in combination with the embodiments.

[0019] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0020] The wireless operation and maintenance in the related art often cannot take into account the saving of computing resources, accuracy, and timeliness of system response for the processing of the detection model. This is determined by the characteristics of the model. The better a model fits a certain type of data, the worse its versatility tends to become. In related academic research, it is usually pursued to improve the accuracy of the model on the basis of ensuring its versatility, but such models are often complex and resource-intensive, and are not suitable for industrial production. Therefore, the related technology either pursues accuracy, subdivides the model in large quantities, consumes resources, and loses scalability when facing new types of data. Or it pursues versatility and uses complex models, thereby facing the problem of increased consumption of inference resources and the inability of the system to respond in time. Both have defects in usability due to large training volume and low accuracy. In response to this problem, the embodiment of the present application is divided into three levels according to the model detection accuracy, namely, the indicator model, the dimensional model, and the object model. The model detection accuracy of the indicator model, the dimensional model, and the object model are ranked in order from low to high. In the embodiment of the present application, the higher the priority of the model detection accuracy, the more accurate the detection result, but the lower the scalability for new types of data. In the actual implementation process, the indicator refers to the key data used to measure the network quality and business busyness during wireless communication, such as connection rate, call drop rate, number of users, etc. The dimension is used to limit the statistical scope of the indicator. Common ones are provinces, cities, communities, etc., and can also be customized user groups. There can be multiple dimensions. The object is the instance pointed to by the true value after the combination of dimension and indicator. For example, if the dimension is province plus city, and the indicator is connection rate, then the connection rate of Nanjing City, Jiangsu Province is an actual object. It should be noted that the embodiment of the present application only divides the detection model into three levels by example. In the actual implementation process, the number of levels for model division can be determined according to the actual situation, and no specific restrictions are made here.

[0021] The embodiments of the present application are aimed at scenarios such as indicator prediction, abnormality warning, root cause analysis (such as fault classification, fault scenario analysis, etc.), and abnormality detection (such as abnormality location, locating the specific root cause of the abnormality through root cause analysis after the abnormality is discovered) in the field of wireless operation and maintenance, and propose a wireless indicator detection method. The operating condition of the wireless communication system and whether there are abnormal status indicators, etc., can be predicted in advance through indicator prediction and abnormality warning, and root cause analysis and abnormality detection can be performed in combination with actual data.

[0022] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a computer terminal as an example, Figure 1 FIG. 1 is a hardware structure block diagram of a computer terminal of a wireless index detection method according to an embodiment of the present invention. Figure 1 As shown, the computer terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the above-mentioned computer terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above-mentioned computer terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0023] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the wireless indicator detection method in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the mobile terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0024] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of a computer terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0025] Figure 2 is a schematic diagram of a wireless indicator detection operation network architecture according to an embodiment of the present invention. The embodiment of the present application can be operated on Figure 2 In the network architecture shown in Figure 2 As shown, the network architecture includes: data management module, data source, model management module, task management module, and front-end management page. The data management module is used to request data from the database for page display and push to the model module for early warning model training and reasoning. The model management module is used to manage the three-layer model and perform training and reasoning tasks. The task management module is used to receive front-end requests, query data from the data module and push it to the model module for processing, as well as manage scheduled tasks. The front-end management page is used to interact with users. Users select the dimensions and indicators that need to be inferred through the front-end page. In addition, users can also set scheduled tasks to continuously update the early warning model.

[0026] In this embodiment, a wireless indicator detection method running on the above-mentioned computer terminal or network architecture is provided. Figure 3 is a flow chart of a wireless indicator detection method according to an embodiment of the present invention. Figure 3 As shown, the process includes the following steps:

[0027] Step S302, selecting the detection model with the highest detection accuracy priority of the existing models as the first detection model for detection according to the detection request;

[0028] In an exemplary embodiment, the hierarchical type of the detection model includes at least one of the following: an indicator model, a dimensional model, and an object model; wherein the model detection accuracy priorities of the indicator model, the dimensional model, and the object model are arranged in order from low to high.

[0029] As described above, the embodiment of the present application is divided into three levels according to the model detection accuracy, namely, the indicator model, the dimension model, and the object model. The priorities of the model detection accuracy of the indicator model, the dimension model, and the object model are arranged from low to high. In the embodiment of the present application, the higher the priority of the model detection accuracy, the more accurate the detection result, but the lower the scalability to new types of data. In the actual implementation process, the number of model layers can be increased or decreased, and is not limited to the three layers of indicators, dimensions, and objects. There can be only indicator and dimension layers or indicator and object layers.

[0030] In an exemplary embodiment, a detection model with the highest priority in model detection accuracy among existing detection models is selected as the first detection model for detection according to a detection request, including: according to the area information and indicator information contained in the detection request, selecting the detection model with the highest priority in model detection accuracy among the existing detection models as the first detection model for detection.

[0031] In the actual implementation process, as mentioned above, the first detection model is one of all existing detection models, and the description of "first" is used only for the purpose of distinction. In the actual implementation process, before the detection is performed, some detection models, namely, existing detection models, already exist. According to the detection request input by the user, among the existing detection models, the detection model with the highest model detection accuracy priority that can match the detection request is selected for detection.

[0032] In an exemplary embodiment, a detection model with the highest priority for model detection accuracy is selected as the first detection model, including: selecting an object model according to indicator information and region information, and if the object model is successfully selected, using the object model as the first detection model for detection; if the object model fails to be selected, selecting a dimensional model according to the region information, and if the dimensional model is successfully selected, using the dimensional model as the first detection model for detection; if the dimensional model fails to be selected, selecting an indicator model according to the indicator information, and using the indicator model as the first detection model for detection.

[0033] In the actual implementation process, the detection request contains regional information and indicator information. When selecting the detection model, the object model is selected first according to the priority order of model detection accuracy. The object model is selected and trained based on the regional information and indicator information, and the corresponding object model needs to be selected based on the regional information and indicator information. In the case of object model selection failure, the matching dimensional model is selected according to the priority order. The dimensional model is trained based on the regional information and needs to be selected based on the regional information. In the case of dimensional model selection failure, the indicator model is selected according to the priority order. The indicator model is trained based on the indicator information and needs to be selected based on the indicator information.

[0034] In the actual implementation process, indicators refer to key data used to measure network quality and business busyness during wireless communication, such as connection rate, call drop rate, number of users, etc. Dimensions are used to limit the statistical scope of indicators. Common ones include province, city, community, etc., and can also be customized user groups. There can be multiple dimensions. Objects are instances pointed to by the real value of the combination of dimensions and indicators. For example, if the dimensions are province plus city, and the indicator is connection rate, then the connection rate of Nanjing City, Jiangsu Province is an actual object.

[0035] In an exemplary embodiment, a detection model with the highest priority in existing model detection accuracy is selected as the first detection model for detection, including: using the first detection model to perform indicator prediction on the wireless indicator corresponding to the detection request to obtain indicator prediction data; using the first detection model to perform early warning detection on the indicator prediction data to obtain early warning detection results.

[0036] In the embodiment of the present invention, after the model detection accuracy in the above steps is obtained and the warning detection result is obtained, an abnormal warning can be performed. In fact, it can also be considered that the warning detection of the indicator prediction data is the abnormal warning of the indicator prediction data.

[0037] In the actual implementation process, the detection model can be divided into a prediction model and an early warning model. The prediction model is used to predict wireless indicators, and the early warning model is used to perform early warning detection on the results of multiple predictions and make abnormal warnings. In the actual implementation process, the detection model can also have the functions of prediction and early warning at the same time, that is, it is not divided into two models, and no specific restrictions are made here. In the actual implementation process, the detection model provided by the embodiment of the present invention can also perform root cause analysis or anomaly detection by detecting actual data, which will not be repeated here.

[0038] Step S304: training a detection model with a higher model detection accuracy priority than the first detection model according to the communication data of the wireless indicator to be detected, so as to detect the next detection request of the same type.

[0039] In an exemplary embodiment, a detection model having a higher model detection accuracy priority than the first detection model is trained according to the communication data of the wireless indicator to be detected, including: when the first detection model is an indicator model, the dimensional model and the object model are trained according to the communication data of the wireless indicator to be detected; when the first detection model is a dimensional model, the object model is trained according to the communication data of the wireless indicator to be detected.

[0040] In the actual implementation process, when the first detection model is the indicator model, it means that when selecting from the existing detection models according to the model detection accuracy priority order, the indicator model is selected, which means that the dimensional model and the object model with a higher model detection accuracy priority than the indicator model are not existing and do not exist. Therefore, it is necessary to train the dimensional model and the object model according to the communication data of the wireless indicator to be detected, in preparation for the next detection request to be detected with the same detection request. When the first detection model is the dimensional model, it means that when selecting from the existing detection models according to the model detection accuracy priority order, the dimensional model is selected, which means that the object model with a higher model detection accuracy priority than the dimensional model does not exist, and it is necessary to train the object model according to the communication data of the wireless indicator to be detected, in preparation for the next detection request to be detected with the same detection request.

[0041] In an exemplary embodiment, the wireless indicator detection method further includes: training and updating the first detection model according to the communication data of the wireless indicator to be detected.

[0042] In the actual implementation process, the existing model is updated and trained to ensure the accuracy of subsequent detection.

[0043] In an exemplary embodiment, before selecting the existing detection model with the highest model detection accuracy priority as the first detection model for detection, the method further includes: acquiring communication data of the wireless indicator to be detected according to the detection request.

[0044] Through the above steps, a wireless indicator detection method is provided, which selects the detection model with the highest priority of model detection accuracy as the first detection model for detection according to the detection request; and trains the detection model with a higher priority of model detection accuracy than the first detection model according to the communication data of the wireless indicator to be detected, so as to detect the next detection request of the same type. This can solve the problem of large training amount and low accuracy of the detection model in the wireless indicator detection in the related technology, and achieves the effect of reducing the training amount of the detection model and improving the accuracy of wireless indicator detection.

[0045] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0046] In this embodiment, a wireless indicator detection device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0047] Figure 4 is a structural block diagram of a wireless indicator detection device according to an embodiment of the present invention. Figure 4 As shown, the detection device 40 includes: a detection selection module 410, which is used to select the detection model with the highest model detection accuracy priority as the first detection type for detection according to the detection request. A detection training module 420 is used to train the detection model with a model detection accuracy priority higher than the first detection model according to the communication data of the wireless indicator to be detected, so as to detect the next detection request of the same type.

[0048] In the actual implementation process, each module in the above-mentioned detection device 40 can also be divided into different functional units to implement each step in the wireless indicator detection method in the above-mentioned embodiment. The specific division and naming of the modules in the detection device 40 are not specifically limited, as long as the various steps in the wireless indicator detection method can be implemented. The specific division of the modules in the detection device 40 is the same as the overall function of the modules in the wireless indicator detection operation network architecture in the above-mentioned embodiment, but the functional division and naming of each individual module can be different, which is no longer specifically limited here.

[0049] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0050] An embodiment of the present invention further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0051] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0052] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0053] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0054] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.

[0055] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0056] In order to enable those skilled in the art to better understand the technical solution of the present invention, it is described below in conjunction with specific scenario embodiments.

[0057] Scenario Example 1

[0058] In this scenario embodiment, the technical solution provided by the embodiment of the present invention is illustrated with examples for indicator prediction and abnormal warning in wireless operation and maintenance detection.

[0059] The wireless indicator detection device provided by the present invention has a core function of dividing the detection model into three-layer models: indicators, dimensions, and objects. The device may initially only provide a model of the indicator layer. When using the device, the user selects indicators under different dimensional combinations for detection. The first dimensional combination will use the indicator model for reasoning to give results to ensure that the operation and maintenance device responds in a timely manner. While reasoning, the data under the dimensional combination will be used to train the corresponding dimensional model and object model. When the indicators under the same dimensional combination are used for reasoning again, the object model that can be matched will be used for reasoning first. For data that does not have a corresponding object model, the dimensional model is used for reasoning, and the data is used to train a new object model. Through the above scheme, the investment in computing resources can be reduced while basically maintaining the accuracy of detection, thereby improving the user experience.

[0060] In the present scenario embodiment, the detection model mentioned in the above embodiment includes a prediction model and an early warning model, wherein the prediction model is used to predict the wireless indicators, and the early warning model is used to warn of abnormalities in the wireless indicators.

[0061] The process steps of the scenario embodiment of the present invention are divided into two parts: abnormal warning model reasoning and abnormal warning model training, both of which can be triggered by the user and performed simultaneously.

[0062] Figure 5 is a schematic diagram of a model reasoning process according to a scenario embodiment of the present invention, such as Figure 5 As shown in the figure, the model reasoning business process includes:

[0063] 1. First, the user selects the dimension and indicator combination that he wants to infer on the front-end page.

[0064] 2. The front end sends an inference request to the task management module.

[0065] 3. The task management module requests query data from the data management module and passes it to the model management module.

[0066] 4. After receiving the data, the model management module detects whether the object model exists. If so, it uses the object model for reasoning.

[0067] 5. If the object model does not exist, continue to check whether the dimensional model exists.

[0068] 6. If a dimensional model exists, use the dimensional model for reasoning. If not, use the indicator model for reasoning.

[0069] 7. The inference results are returned to the front end for display.

[0070] Figure 6 is a schematic diagram of the detection algorithm reasoning process according to a scenario embodiment of the present invention, such as Figure 6 As shown in the figure, the prediction model is first used to predict the data for the next n days, and then the early warning model is used to detect whether there are anomalies in the predicted data. If there are anomalies, they are output as early warning results. The prediction model and early warning model used in the process can be replaced by various prediction algorithms and early warning algorithms.

[0071] Figure 7 Schematic diagram of the model training process according to the scenario embodiment of the present invention. Figure 7 As shown, including:

[0072] 1. First, the user selects the dimension and indicator combination that he wants to infer on the front-end page.

[0073] 2. The front end sends an inference request to the task management module.

[0074] 3. The task management module requests query data from the data management module and passes it to the model management module.

[0075] 4. After receiving the data, the model management module detects whether there are corresponding models at each layer.

[0076] 5. If the model exists, update the old model.

[0077] 6. If the model does not exist, train a new model.

[0078] Figure 8 is a schematic diagram of the detection algorithm training process according to a scenario embodiment of the present invention, such as Figure 8 As shown, in the scenario embodiment of the present invention, the prediction model can be trained first, and then the early warning model can be trained. The prediction model and early warning model used in the process can be replaced by various prediction algorithms and early warning algorithms.

[0079] Scenario Example 2

[0080] The process of abnormal warning of the number of VoLTE users in the city dimension on a newly deployed system is as follows:

[0081] 1. Users select dimensions and indicators on the front-end page to query data.

[0082] 2. Select the abnormal warning function for reasoning.

[0083] At this time, the system is in a brand new state, and only the indicator model of the number of VoLTE users exists. According to the process of the embodiment of the present invention, the indicator model will be selected for reasoning at this time.

[0084] 3. Backstage synchronous training model.

[0085] When using the indicator model of the number of VoLTE users for reasoning, the background will simultaneously train new dimensional models and object models, as well as update the existing indicator models. The dimensional model at this time is a model of the city dimension, that is, all VoLTE user number indicators with the city dimension can use this model for reasoning. The object model generated by training is in the city dimension, which can be 7 different models from Xuzhou to Zhenjiang (for example only, without specific limitations on the solution of the present invention). To use this model, the corresponding object model can be used only when the indicator is the number of VoLTE users, the dimension is the city, and the dimension value is Zhenjiang (one of the 7 cities).

[0086] 4. The inference results are returned to the front end for display.

[0087] Scenario Example 3

[0088] Based on the training results of scenario embodiment 2, in scenario embodiment 3, a new object is selected for abnormal warning.

[0089] 1. Select dimensions and indicators to query data.

[0090] In this scenario embodiment, the selected dimension and indicator is the number of VoLTE users in Nanjing.

[0091] 2. Select the abnormal warning function for reasoning.

[0092] At this time, since the city-dimensional VoLTE user number model has been trained in the second scenario implementation example, but there is no object-dimensional model for Nanjing, the former dimensional model is preferentially selected for reasoning.

[0093] 3. Backstage synchronous training model.

[0094] When using the VoLTE user number indicator model for inference, the background will simultaneously train a new object model and update the existing indicator model and dimension model.

[0095] 4. The inference results are returned to the front end for display.

[0096] In the actual implementation process, when the operation and maintenance scenario is a large screen guarantee, such as an application scenario where the dimensions are fixed or not concerned with the dimensions, the dimension model can usually be deleted, because no new dimensions will appear at this time. Or when not paying attention to specific objects but only to overall indicators, the object model can be deleted, leaving only the indicators and dimensions.

[0097] It should be noted that in the scenario embodiments of the present invention (including scenario embodiments one, two, and three), only the steps of indicator prediction and abnormal warning are described. In the actual implementation process, the wireless indicator detection device provided in the scenario embodiments of the present invention is used to execute the wireless indicator detection method in the above embodiment. In addition to being applicable to indicator prediction and abnormal warning in the field of wireless operation and maintenance, it can also be applicable to scenarios of root cause analysis and abnormality detection.

[0098] The present invention relates to a wireless indicator detection method based on a hierarchical model, and in particular to network operation and maintenance in the field of communications. Network operation and maintenance usually evaluates network conditions and locates network problems by monitoring and analyzing various quantitative indicators generated during the communication process. Abnormal warning is to identify the possibility of abnormalities in a unit in the operation and maintenance system in advance through an algorithm to avoid problems. The current network operation and maintenance system is developing in the direction of intelligent operation and maintenance. The embodiment of the present invention aims to provide a wireless indicator detection device based on model layering to improve the intelligence and availability of the operation and maintenance system. The present invention proposes a wireless indicator detection technology based on model layering, first providing a batch of models with strong versatility and poor accuracy to meet the timeliness of system response, and then training new models with poor versatility but high accuracy during use, and only training the models that need to be used to avoid waste of resources.

[0099] The technical solution of the embodiment of the present invention can be applied to the operation and maintenance system of wireless indicators under 4 / 5G communication networks, and is oriented to the indicator prediction, abnormal warning, root cause analysis, and abnormal detection scenarios of wireless indicators. Communication operators all have their own big data operation and maintenance systems, and they also have the need to develop in the direction of intelligent operation and maintenance. The present invention only needs simple adaptation to be embedded in various operation and maintenance systems, and can also be redeveloped using the solution of the present invention. The detection algorithm in the embodiment of the present invention can be replaced by various prediction and analysis algorithms, etc. The indicator detection method of the present invention that first predicts indicators and then warns of abnormalities, the prediction and detection steps can also be replaced by various prediction and detection algorithms.

[0100] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A wireless index detection method, characterized in that: include: According to the detection request, a detection model with the highest detection accuracy priority among the existing models is selected as the first detection model for detection; According to the communication data of the wireless indicator to be detected, a detection model with a higher model detection accuracy priority than the first detection model is trained to detect the next detection request of the same type.

2. The method according to claim 1, characterized in that The hierarchical type of the detection model includes at least one of the following: Indicator model, dimensional model, object model; wherein the model detection accuracy priorities of the indicator model, the dimensional model, and the object model are arranged in descending order.

3. The method according to claim 2, characterized in that The step of selecting, according to the detection request, a detection model with the highest detection accuracy priority of an existing model as a first detection model for detection includes: According to the area information and indicator information included in the detection request, the detection model with the highest model detection accuracy priority is selected from the existing detection models as the first detection model for detection.

4. The method according to claim 3, characterized in that The step of selecting the detection model with the highest model detection accuracy priority as the first detection model includes: selecting the object model according to the indicator information and the area information, and if the object model is successfully selected, using the object model as the first detection model for detection; In the case where the object model fails to be selected, the dimensional model is selected according to the region information, and in the case where the dimensional model is successfully selected, the dimensional model is used as the first detection model for detection; In the case where the dimensional model fails to be selected, the indicator model is selected according to the indicator information, and the indicator model is used as the first detection model for detection.

5. The method according to claim 1, characterized in that The step of selecting the detection model with the highest detection accuracy priority of the existing models as the first detection model for detection includes: Using the first detection model to perform indicator prediction on the wireless indicator corresponding to the detection request to obtain indicator prediction data; The first detection model is used to perform early warning detection on the indicator prediction data to obtain an early warning detection result.

6. The method according to claim 2, characterized in that The step of training, according to the communication data of the wireless indicator to be detected, a detection model having a higher priority in model detection accuracy than the first detection model, comprises: In the case where the first detection model is the indicator model, training the dimensional model and the object model according to the communication data of the wireless indicator to be detected; In the case where the first detection model is the dimensional model, the object model is trained according to the communication data of the wireless indicator to be detected.

7. The method according to claim 1, characterized in that Also includes: The first detection model is trained and updated according to the communication data of the wireless indicator to be detected.

8. The method according to claim 1, characterized in that Before selecting the detection model with the highest detection accuracy priority of the existing models as the first detection model for detection, the method further includes: The communication data of the wireless indicator to be detected is obtained according to the detection request.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program implements the method described in any one of claims 1 to 8 when executed by a processor.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method described in any one of claims 1 to 8 is implemented.

Images