Method and device for predicting MOS value in communication scene

Abstract

The invention provides a method and device for predicting an MOS value in a communication scene, and relates to the technical field of communication. The scheme comprises the following steps: after real-time alarm information corresponding to a current network element related to an average opinion score MOS value is monitored, analyzing the real-time alarm information to extract a target cell identifier carried by the real-time alarm information; obtaining N pieces of latest historical alarm information related to the target cell identifier and including the real-time alarm information; constructing an alarm information time sequence based on the latest N pieces of historical alarm information; the alarm information time sequence is input into a trained MOS value prediction model, an MOS predicted value corresponding to the target cell identifier output by the MOS value prediction model is obtained, and the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model. On-site measurement is not needed, so that the cost is greatly reduced; the dynamic updating of the MOS value is realized, and higher real-time performance is provided; all network cells can be covered, and overall monitoring of network performance is ensured.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a method and device for predicting MOS values ​​in a communication scenario. Background Art

[0002] In the process of network maintenance and optimization, the Mean Opinion Score (MOS) value is an important indicator for measuring voice or video quality and is widely used in network quality assessment. The MOS value converts voice or video quality into a value from 1 to 5 through subjective testing, where 5 represents the best quality and 1 represents the worst quality. This indicator can truly reflect the user's perceived experience, so it is crucial in network quality management.

[0003] The traditional MOS collection method relies on carrying special equipment and transmitting and receiving voice or video through the actual network for testing. This method usually involves testing each cell and each time period, and analyzing the data through quality assessment software to obtain the MOS value. Although this method can provide detailed network quality data, its characteristics include high human and material investment, and in remote mountainous areas, the collection can only be performed once when the network is opened, and it is almost never updated afterwards, resulting in the inability to monitor the changes in MOS values ​​in real time. The collection of MOS values ​​for all cells in the network faces huge challenges, and a low-cost, timely and comprehensive solution is needed to overcome the limitations of traditional methods. Summary of the invention

[0004] The present disclosure provides a method and device for predicting MOS values ​​in a communication scenario, so as to at least solve the problems of high cost and delay in related technologies.

[0005] The first aspect of the present application proposes a method for predicting a MOS value in a communication scenario, comprising: after monitoring real-time alarm information corresponding to an existing network element related to a mean opinion score (MOS) value, parsing the real-time alarm information to extract a target cell identifier carried by the real-time alarm information; obtaining the most recent N historical alarm information including the real-time alarm information related to the target cell identifier, wherein the N historical alarm information are all alarm information related to the MOS value; constructing an alarm information time series based on the most recent N historical alarm information; inputting the alarm information time series into a trained MOS value prediction model, and obtaining a MOS prediction value corresponding to the target cell identifier output by the MOS value prediction model, wherein the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model.

[0006] According to one embodiment of the present application, the most recent N historical alarm information including real-time alarm information related to the target cell identifier is obtained, including: extracting the historical alarm information related to the target cell identifier from an alarm information database; tracing back the historical alarm information in chronological order starting from the real-time alarm information, and obtaining the most recent N historical alarm information including the real-time alarm information.

[0007] According to one embodiment of the present application, an alarm information time series is constructed based on the most recent N historical alarm messages, including: extracting key fields from each of the most recent N historical alarm messages to obtain key field information corresponding to each of the most recent N historical alarm messages; and constructing an alarm information time series based on the key field information corresponding to each of the most recent N historical alarm messages.

[0008] According to one embodiment of the present application, a training method for a MOS value prediction model includes: obtaining multiple groups of drive test data, wherein each group of drive test data includes 1 drive test cell identifier, 1 drive test collection time, and 1 drive test MOS actual value; for any group of drive test data, obtaining historical alarm information samples corresponding to the drive test cell identifier included in the group of drive test data, and tracing back the historical alarm information samples from the drive test collection time included in the group of drive test data to obtain N traced historical alarm information samples; for any group of drive test data, constructing training samples according to the group of drive test data and the N historical alarm information samples corresponding to the group of drive test data; obtaining training samples corresponding to each of the multiple groups of drive test data to form a training sample set, and training an initial MOS value prediction model based on the training sample set until the training is completed, and obtaining the MOS value prediction model generated by the training.

[0009] According to an embodiment of the present application, for any group of drive test data, a training sample is constructed based on the group of drive test data and N historical alarm information samples corresponding to the group of drive test data, including: for any group of drive test data, an alarm information time series sample is constructed based on the N historical alarm information samples corresponding to the group of drive test data, and the actual value of the drive test MOS included in the group of drive test data is used as the sample label of the alarm information time series sample; for any group of drive test data, a training sample is composed based on the alarm information time series sample and the sample label corresponding to the group of drive test data.

[0010] According to an embodiment of the present application, for any group of drive test data, a training sample is constructed based on the group of drive test data and N historical alarm information samples corresponding to the group of drive test data, including: for any group of drive test data, an alarm information time series sample is jointly constructed based on the drive test cell identifier, the drive test collection time and the N historical alarm information samples corresponding to the group of drive test data included in the group of drive test data, and the drive test cell identifier and the actual value of the drive test MOS included in the group of drive test data are used as sample labels of the alarm information time series sample; for any group of drive test data, a training sample is formed based on the alarm information time series sample and the sample label corresponding to the group of drive test data.

[0011] According to one embodiment of the present application, the method for predicting the MOS value in a communication scenario further includes: comparing the MOS prediction value corresponding to the target cell identifier with a preset threshold to obtain a first comparison result; in response to the first comparison result indicating that the MOS prediction value corresponding to the target cell identifier is less than the preset threshold, sending a road test instruction and / or a network maintenance instruction to the terminal device of the maintenance personnel corresponding to the target cell identifier.

[0012] According to one embodiment of the present application, a method for predicting a MOS value in a communication scenario further includes: in response to a first comparison result indicating that the MOS prediction value corresponding to the target cell identifier is less than a preset threshold, obtaining a model output time of the MOS prediction value corresponding to the target cell identifier; determining a target time period formed by tracing back a preset time length from the model output time; obtaining at least one target user number related to the target cell identifier within the target time period; and sending a preset text to the target user number.

[0013] According to one embodiment of the present application, a method for predicting a MOS value in a communication scenario, the method also includes: obtaining a previous MOS prediction value corresponding to a target cell identifier; comparing the previous MOS prediction value corresponding to the target cell identifier with the MOS prediction value corresponding to the target cell identifier at the current moment to obtain a second comparison result; in response to the second comparison result indicating that the previous MOS prediction value corresponding to the target cell identifier and the MOS prediction value corresponding to the target cell identifier at the current moment are different in value, obtaining a target color corresponding to the MOS prediction value corresponding to the target cell identifier at the current moment; and updating the color of a target position corresponding to the target cell identifier on a geographic information system GIS map based on the target color.

[0014] The second aspect of the present application proposes a device for predicting MOS values ​​in a communication scenario, including: a monitoring module, which is used to parse the real-time alarm information to extract the target cell identifier carried by the real-time alarm information after monitoring the real-time alarm information corresponding to the existing network element related to the mean opinion score MOS value; an acquisition module, which is used to obtain the latest N historical alarm information related to the target cell identifier, including the real-time alarm information, wherein the N historical alarm information are all alarm information related to the MOS value; a construction module, which is used to construct an alarm information time series based on the latest N historical alarm information; a prediction module, which is used to input the alarm information time series into a trained MOS value prediction model, and obtain the MOS prediction value corresponding to the target cell identifier output by the MOS value prediction model, wherein the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model.

[0015] According to one embodiment of the present application, the acquisition module is also used to: extract historical alarm information related to the target cell identifier from the alarm information database; trace back the historical alarm information in chronological order starting from the real-time alarm information, and obtain the most recent N historical alarm information including the real-time alarm information.

[0016] According to one embodiment of the present application, the construction module is also used to: extract key fields from each of the most recent N historical alarm information to obtain key field information corresponding to each of the most recent N historical alarm information; and construct an alarm information time series based on the key field information corresponding to each of the most recent N historical alarm information.

[0017] According to one embodiment of the present application, the device for predicting MOS values ​​in a communication scenario further includes a training module, which is used to: according to one embodiment of the present application, obtain multiple groups of drive test data, wherein each group of drive test data includes 1 drive test cell identifier, 1 drive test collection time and 1 drive test MOS actual value; for any group of drive test data, obtain historical alarm information samples corresponding to the drive test cell identifier included in the group of drive test data, and trace back the historical alarm information samples from the drive test collection time included in the group of drive test data to obtain N traced historical alarm information samples; for any group of drive test data, construct training samples according to the group of drive test data and the N historical alarm information samples corresponding to the group of drive test data; obtain training samples corresponding to each of the multiple groups of drive test data to form a training sample set, and train the initial MOS value prediction model based on the training sample set until the training is completed, and obtain the MOS value prediction model generated by the training.

[0018] According to one embodiment of the present application, the training module is also used to: for any group of road test data, construct an alarm information time series sample based on N historical alarm information samples corresponding to the group of road test data, and use the actual value of the road test MOS included in the group of road test data as the sample label of the alarm information time series sample; for any group of road test data, form a training sample based on the alarm information time series sample and the sample label corresponding to the group of road test data.

[0019] According to one embodiment of the present application, the training module is further used to: for any group of drive test data, jointly construct an alarm information time series sample based on the drive test cell identifier, the drive test collection time and N historical alarm information samples corresponding to the group of drive test data included in the group of drive test data, and use the drive test cell identifier and the actual value of the drive test MOS included in the group of drive test data as sample labels of the alarm information time series sample; for any group of drive test data, form a training sample based on the alarm information time series sample and the sample label corresponding to the group of drive test data.

[0020] According to one embodiment of the present application, the device for predicting the MOS value in a communication scenario also includes a post-processing module, which is used to: compare the MOS prediction value corresponding to the target cell identifier with a preset threshold to obtain a first comparison result; in response to the first comparison result indicating that the MOS prediction value corresponding to the target cell identifier is less than the preset threshold, send a road test instruction and / or a network maintenance instruction to the terminal device of the maintenance personnel corresponding to the target cell identifier.

[0021] According to one embodiment of the present application, the post-processing module is further used to: in response to the first comparison result indicating that the MOS prediction value corresponding to the target cell identifier is less than a preset threshold, obtain the model output time of the MOS prediction value corresponding to the target cell identifier; determine a target time period formed by tracing back a preset time length from the model output time; obtain at least one target user number related to the target cell identifier within the target time period; and send a preset text to the target user number.

[0022] According to one embodiment of the present application, the post-processing module is also used to: obtain the previous MOS prediction value corresponding to the target cell identifier; compare the previous MOS prediction value corresponding to the target cell identifier with the MOS prediction value corresponding to the target cell identifier at the current moment to obtain a second comparison result; in response to the second comparison result indicating that the previous MOS prediction value corresponding to the target cell identifier and the MOS prediction value corresponding to the target cell identifier at the current moment are different in value, obtain the target color corresponding to the MOS prediction value corresponding to the target cell identifier at the current moment; and update the color of the target position corresponding to the target cell identifier on the geographic information system GIS map based on the target color.

[0023] The third aspect embodiment of the present application proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to implement the method for predicting MOS values ​​in a communication scenario as described in the first aspect embodiment of the present application.

[0024] The fourth aspect embodiment of the present application proposes a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to implement the method for predicting MOS values ​​in a communication scenario as described in the first aspect embodiment of the present application.

[0025] The fifth aspect embodiment of the present application proposes a computer program product, including a computer program, which, when executed by a processor, implements the method for predicting MOS values ​​in a communication scenario as described in the first aspect embodiment of the present application.

[0026] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects:

[0027] The present application uses existing alarm information in the network to predict MOS values, eliminating the need for on-site measurements, thereby greatly reducing costs.

[0028] The prediction of the MOS value of this application is based on real-time alarm information, and the update is triggered by the alarm information, thereby realizing the dynamic update of the MOS value. Compared with the traditional quarterly or monthly road test MOS collection method, this method provides higher real-time performance.

[0029] Traditional drive tests usually can only cover a limited area. This application uses alarm information to perform MOS prediction, which can cover all network cells and ensure overall monitoring of network performance.

[0030] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0032] Figure 1 It is a schematic diagram of an exemplary implementation of a method for predicting MOS value in a communication scenario shown in an embodiment of the present application.

[0033] Figure 2 It is a schematic diagram showing an embodiment of the present application for obtaining the latest N historical alarm information including real-time alarm information related to the target cell identifier to predict the MOS value.

[0034] Figure 3 It is a schematic diagram of an exemplary implementation of a method for predicting MOS value in a communication scenario shown in an embodiment of the present application.

[0035] Figure 4 It is a schematic diagram of an exemplary implementation of a method for predicting MOS value in a communication scenario shown in an embodiment of the present application.

[0036] Figure 5 It is a schematic diagram of a device for predicting MOS value in a communication scenario shown in an embodiment of the present application.

[0037] Figure 6 It is a schematic diagram of an electronic device shown in one embodiment of the present application. DETAILED DESCRIPTION

[0038] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0039] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations.

[0040] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0041] The following is an introduction to some professional terms involved in this application:

[0042] MOS (Mean Opinion Score) is a commonly used subjective evaluation indicator used to evaluate the quality of communication services. ( Qua li t y o fS erv i ce , Q o S ) 。 MOS Values ​​are often used to measure the quality of voice and video communications.

[0043] The voice MOS value is an indicator used to evaluate the quality of voice communication. It is usually based on the user's subjective evaluation of the voice call quality, ranging from 1 to 5. 1 represents extremely poor voice quality, and 5 represents excellent voice quality.

[0044] The video MOS value is an indicator used to evaluate the quality of video communication. It is also based on the user's subjective evaluation of video quality, and the range is usually from 1 to 5. 1 represents extremely poor video quality, and 5 represents excellent video quality.

[0045] Figure 1 is a schematic diagram of an exemplary implementation of a method for predicting MOS values ​​in a communication scenario shown in the present application, such as Figure 1 As shown, the method for predicting the MOS value in the communication scenario includes the following steps:

[0046] S101, after monitoring the real-time alarm information corresponding to the existing network element related to the mean opinion score MOS value, the real-time alarm information is parsed to extract the target cell identifier carried in the real-time alarm information.

[0047] The operation of existing network elements, such as the mobile station (MS), base station (BS), transmission resource adapter (TRA), media gateway (MGW), IP multimedia subsystem (IMS), etc., plays a vital role in the network. Their operation status directly affects the quality of service (QoS), especially the mean opinion score (MOS) value. The operation status of the network element can be reflected through the network element alarm information. The idea of ​​this application is to predict the MOS value of each cell based on the alarm information of the existing network element.

[0048] In this application, the alarm situation of the entire network can be monitored in real time, the original alarm information flow can be filtered, and only the focus alarm can be identified and filtered out. Subsequently, N historical alarm information can be traced back based on the focus alarm. A focus alarm is an alarm that will cause the model to calculate and update the MOS value, and can also be understood as an alarm related to the MOS value. After monitoring the real-time alarm information corresponding to the existing network element related to the MOS value (which is a focus alarm), the real-time alarm information is parsed to extract the target cell identifier carried by the real-time alarm information.

[0049] In the initial stage, the present invention sets the alarms of the cell type as focus alarms. For the alarms of other types of network elements, the model will ignore these alarms and will not include them in the calculation and update of the MOS value (MeanOpinion Score). This strategy is intended to simplify the complexity of the model and ensure that the model can run efficiently with limited computing resources.

[0050] As computing power gradually increases, the scope of focus alarms will be gradually expanded in the future. This is because in theory, there is a correlation between alarms of different network element types and they may also cause changes in MOS values. For example, a transmission interruption may cause a cell failure, so there is a causal relationship between the transmission interruption alarm and the cell alarm. By expanding the scope of focus alarms, the actual state of the network can be more accurately reflected, and the accuracy of MOS value prediction can be improved.

[0051] From the perspective of network elements, in the initial stage, only cell alarms corresponding to specific MOS values ​​are selected to train the model. This strategy ensures that the model can focus on key factors that directly affect the MOS value in the initial stage. Later, it can be considered to expand the scope of network elements, because the network elements that affect the MOS value are not limited to the radio access network (RAN), but also include the core network (CoreNetwork), the transport network (Transport Network) and the bearer network (Backhaul Network). In the future, try to use all network elements and alarms of the entire network to train the model to more comprehensively reflect the overall status of the network.

[0052] From the perspective of alarm status, the current status of network devices will inevitably affect the MOS value. When an alarm is generated, the MOS value may decrease; when the alarm is restored, the MOS value may increase. Therefore, both active alarms (ActiveAlarms) and cleared alarms (ClearedAlarms) should be considered within the selected data range. By comprehensively considering the generation and restoration status of alarms, the impact of the status changes of network devices on the MOS value can be more accurately evaluated.

[0053] S102: Acquire the latest N pieces of historical alarm information including real-time alarm information related to the target cell identifier.

[0054] It is not a single alarm message that affects the MOS value, but a combination of a series of alarm messages. These alarms are not only independent, but also have obvious time characteristics. For example, some alarm messages have a mutually derived relationship in the time dimension, that is, the occurrence of some alarms may trigger other alarms. In addition, over time, active alarms (ActiveAlarms) and cleared alarms (ClearedAlarms) jointly reflect the changes in cell quality, and these changes will affect the MOS value (Mean Opinion Score) of the cell.

[0055] In the present application, after the target cell identifier is determined, the latest N historical alarm information including real-time alarm information related to the target cell identifier is obtained to perform the next step of MOS value prediction. Optionally, N can be set to 100.

[0056] The N pieces of historical alarm information are all alarm information related to the MOS value, that is, they are all focus alarms.

[0057] S103: construct an alarm information time series based on the latest N pieces of historical alarm information.

[0058] As a feasible approach, a time series of alarm information is directly constructed based on N pieces of historical alarm information.

[0059] It is understandable that the original alarm information of the existing network contains many fields, and each alarm record usually contains 118 fields. Among them, some fields have overlapping meanings, resulting in information redundancy, and some fields are irrelevant to the operating status of the network element, and therefore have no direct impact on the change of the MOS value.

[0060] Therefore, in order to reduce the complexity of subsequent models and improve processing efficiency, as another feasible method, key fields are extracted for each of the most recent N historical alarm messages, the key field information corresponding to each of the most recent N historical alarm messages is obtained, and an alarm information time series is constructed based on the key field information corresponding to each of the most recent N historical alarm messages.

[0061] Exemplarily, the key fields are set to message ID, network element ID, network element type, positioning object, positioning object type, event occurrence time, clearing time, first-level major, second-level major, standardized alarm category, clearing status, warning flag, alarm event level, and equipment manufacturer, a total of 14 fields.

[0062] It is not difficult to understand that the key fields can be set according to actual conditions. The 14 key fields illustrated in this application are only examples and do not constitute a limitation to this application.

[0063] If a field of a piece of historical alarm information is empty, the field is set to a default value to indicate that the field of the piece of historical alarm information is empty.

[0064] S104, inputting the alarm information time series into a trained MOS value prediction model, and obtaining a MOS prediction value corresponding to the target cell identifier output by the MOS value prediction model, wherein the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model.

[0065] The MOS value prediction model in this application uses the Long Short-Term Memory (LSTM) network for data fitting. Compared with other types of neural networks, the LSTM network is specially designed to solve long-term dependency problems and has strong memory capabilities. LSTM can effectively capture and maintain dependencies in long sequences through its unique gating mechanism, which makes it perform well in processing time series data.

[0066] Exemplarily, if the MOS value prediction model is a voice MOS value prediction model, the alarm information time series is input into the trained voice MOS value prediction model to obtain the voice MOS prediction value corresponding to the target cell identifier output by the voice MOS value prediction model.

[0067] Exemplarily, if the MOS value prediction model is a video MOS value prediction model, the alarm information time series is input into a trained video MOS value prediction model to obtain a video MOS prediction value corresponding to a target cell identifier output by the video MOS value prediction model.

[0068] For ease of understanding, Figure 2 This is a schematic diagram of obtaining the latest N historical alarm information including real-time alarm information related to the target cell identifier to predict the MOS value, as shown in the present application. Figure 2 As shown, after monitoring the latest 1 real-time alarm information corresponding to the existing network element related to the MOS value, the latest N pieces of real-time alarm information including the real-time alarm information related to the target cell identifier carried by it are obtained ( Figure 2 (Take N as 100 for example) The historical alarm information is predicted by inputting the MOS value prediction model to obtain the MOS prediction value corresponding to the latest real-time alarm information.

[0069] The embodiment of the present application proposes a method for predicting a MOS value in a communication scenario, comprising: after monitoring real-time alarm information corresponding to an existing network element related to a mean opinion score (MOS) value, parsing the real-time alarm information to extract a target cell identifier carried by the real-time alarm information; obtaining the latest N historical alarm information related to the target cell identifier, including the real-time alarm information; constructing an alarm information time series based on the latest N historical alarm information; inputting the alarm information time series into a trained MOS value prediction model, and obtaining a MOS prediction value corresponding to the target cell identifier output by the MOS value prediction model, wherein the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model.

[0070] The present application uses existing alarm information in the network to predict MOS values, eliminating the need for on-site measurements, thereby greatly reducing costs.

[0071] The prediction of the MOS value of this application is based on real-time alarm information, and the update is triggered by the alarm information, thereby realizing the dynamic update of the MOS value. Compared with the traditional quarterly or monthly road test MOS collection method, this method provides higher real-time performance.

[0072] Traditional drive tests usually can only cover a limited area. This application uses alarm information to perform MOS prediction, which can cover all network cells and ensure overall monitoring of network performance.

[0073] Among them, in S102, the latest N historical alarm information including real-time alarm information related to the target cell identifier is obtained, including: extracting the historical alarm information related to the target cell identifier from the alarm information database; tracing back the historical alarm information in chronological order starting from the real-time alarm information, and obtaining the latest N historical alarm information including the real-time alarm information.

[0074] Figure 3 is a schematic diagram of an exemplary implementation of a method for predicting MOS values ​​in a communication scenario shown in the present application, such as Figure 3 As shown, the method for predicting the MOS value in the communication scenario includes the following steps:

[0075] S301, obtaining multiple groups of drive test data, wherein each group of drive test data includes a drive test cell identifier, a drive test collection time, and a drive test MOS actual value.

[0076] The drive test cell identifier (CELL_ID) is a unique identifier used to indicate a drive test cell.

[0077] The drive test collection time records the specific time when the group of drive test data is collected.

[0078] The actual drive test MOS value refers to the actual MOS value of the cell corresponding to the drive test cell identifier collected during the drive test collection time. It should be noted that if the voice MOS value prediction model is trained, the actual voice MOS value is collected here, and if the video MOS value prediction model is trained, the actual video MOS value is collected here.

[0079] For example, a set of drive test data may be recorded as: CELL_ID1, 2024-08-01 09:00, voice MOS value 1. This means that the voice MOS value collected at 09:00 on August 1, 2024 in the cell with the drive test cell ID 1 is 1.

[0080] S302, for any group of drive test data, obtain historical alarm information samples corresponding to the drive test cell identifier included in the group of drive test data, and trace back the historical alarm information samples from the drive test collection time included in the group of drive test data to obtain N traced historical alarm information samples.

[0081] Exemplarily, if the drive test data is: CELL_ID1, 09:00 on August 1, 2024, voice MOS value 1. The historical alarm information corresponding to the cell with the drive test cell ID 1 is obtained as the historical alarm information sample, and according to the collection timestamps of all historical alarm information samples corresponding to the cell with the drive test cell ID 1, the N traced back historical alarm information samples are obtained starting from 09:00 on August 1, 2024.

[0082] S303: For any group of drive test data, construct a training sample according to the group of drive test data and N historical alarm information samples corresponding to the group of drive test data.

[0083] The first training sample construction method:

[0084] For any set of drive test data, an alarm information time series sample is constructed according to the N historical alarm information samples corresponding to the set of drive test data (key field extraction is required), and the actual drive test MOS value included in the set of drive test data is used as the sample label of the alarm information time series sample.

[0085] For any set of drive test data, after obtaining the alarm information time series samples and sample labels of the alarm information time series samples corresponding to the set of drive test data, training samples are formed based on the alarm information time series samples and sample labels corresponding to the set of drive test data.

[0086] In this way, if there are 10,000 sets of drive test data, then 10,000 corresponding training samples are obtained.

[0087] The second training sample construction method:

[0088] For any group of drive test data, an alarm information time series sample is jointly constructed based on the drive test cell identifier, drive test collection time and N historical alarm information samples corresponding to the group of drive test data (key field extraction is also required), and the drive test cell identifier and the actual value of the drive test MOS included in the group of drive test data are used as sample labels of the alarm information time series sample.

[0089] For any set of drive test data, after obtaining the alarm information time series samples and sample labels of the alarm information time series samples corresponding to the set of drive test data, training samples are formed based on the alarm information time series samples and sample labels corresponding to the set of drive test data.

[0090] Similarly, if there are 10,000 sets of road test data, then 10,000 corresponding training samples will be obtained.

[0091] Compared with the first training sample construction method, the second training sample construction method adds the drive test cell identification and the drive test collection time as training features, making the trained MOS value prediction model more accurate.

[0092] Taking the second training sample construction method as an example, the drive test cell identifier, the drive test collection time, and the N historical alarm information samples corresponding to the group of drive test data together form a matrix data, which is a time series sample of the alarm information and can be expressed as:

[0093] CELL_ID1, drive test collection time 1, historical alarm information 1 (14 key fields are extracted, including message identifier, network element identifier, network element type, positioning object, positioning object type, event occurrence time, clearing time, first-level major, second-level major, standardized alarm category, clearing status, warning flag, alarm (event) level, equipment manufacturer);

[0094] CELL_ID1, drive test collection time 1, historical alarm information 2 (14 key fields are extracted, including message identifier, network element identifier, network element type, positioning object, positioning object type, event occurrence time, clearing time, first-level major, second-level major, standardized alarm category, clearing status, warning flag, alarm (event) level, equipment manufacturer); ............

[0096] CELL_ID1, road test collection time 1, historical alarm information N (14 key fields are extracted, including message identifier, network element identifier, network element type, positioning object, positioning object type, event occurrence time, clearing time, first-level major, second-level major, standardized alarm category, clearing status, warning flag, alarm (event) level, equipment manufacturer).

[0097] S304, obtaining training samples corresponding to each of the multiple groups of drive test data to form a training sample set, and training the initial MOS value prediction model based on the training sample set until the training is completed, and obtaining the MOS value prediction model generated by the training.

[0098] If the training samples constructed by the first training sample construction method are used for training, the number of input features of the MOS value prediction model is 14 (14 key fields); the number of output features is 1 (MOS prediction value); and the input sequence length is N.

[0099] If the training samples constructed by the second training sample construction method are used for training, the number of input features of the MOS value prediction model is 16 (1 drive test cell identifier, 1 drive test collection time, and 14 key fields); the number of output features is 2 (drive test cell identifier, MOS prediction value); and the input sequence length is N.

[0100] Furthermore, the number of hidden layer units of the MOS value prediction model is set to 64, and each hidden layer contains 64 LSTM units.

[0101] Furthermore, the activation function of the MOS value prediction model adopts Softmax.

[0102] Furthermore, the loss function of the MOS value prediction model uses categorical_crossentropy. Categorical_crossentropy is suitable for multi-classification problems. When the model output passes through the Softmax activation function, the categorical cross entropy can evaluate the difference between the predicted probability distribution and the true label distribution.

[0103] This application uses model.fit(X_train, y_train, epochs = 1, batch_size = 1, verbose = 1) # training model. Here, for each training round, the batch size is 1 (that is, 1 sample is input each time), and detailed information is output (verbose = 1).

[0104] model.fit is a fitting function used to train a neural network model. By training the model with training data, the model parameters that match the training data can be fitted so that the model can predict or classify new input data. By iterating the training data multiple times, the fit function can automatically adjust the model parameters to minimize the loss function, thereby obtaining a model with good performance.

[0105] S305, after monitoring the real-time alarm information corresponding to the existing network element related to the mean opinion score MOS value, the real-time alarm information is parsed to extract the target cell identifier carried in the real-time alarm information.

[0106] S306: Obtain the latest N pieces of historical alarm information including real-time alarm information related to the target cell identifier.

[0107] S307: construct an alarm information time series based on the latest N pieces of historical alarm information.

[0108] S308, inputting the alarm information time series into a trained MOS value prediction model, and obtaining a MOS prediction value corresponding to the target cell identifier output by the MOS value prediction model, wherein the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model.

[0109] The embodiment of the present application uses the existing alarm information in the network to predict the MOS value, without the need for on-site measurement, thereby greatly reducing costs; the prediction of the MOS value of the present application is based on real-time alarm information, and the update is triggered by the alarm information, thereby realizing the dynamic update of the MOS value. Compared with the traditional quarterly or monthly road test MOS collection method, this method provides higher real-time performance; traditional road tests usually can only cover a limited area, and the present application uses alarm information to predict MOS, which can cover all network cells and ensure the overall monitoring of network performance.

[0110] Figure 4 is a schematic diagram of an exemplary implementation of a method for predicting MOS values ​​in a communication scenario shown in the present application, such as Figure 4 As shown, the method for predicting the MOS value in the communication scenario includes the following steps:

[0111] Monitor the alarm situation of the entire network in real time, pre-process the real-time alarm flow of the existing network, and only filter out the alarms related to the MOS value. After monitoring the real-time alarm information corresponding to the existing network element related to the MOS value, parse the real-time alarm information to extract the target cell identifier carried by the real-time alarm information, and obtain the latest N historical alarm information related to the target cell identifier, including the real-time alarm information (the N historical alarm information are all alarms related to the MOS value).

[0112] The key fields of the most recent N historical alarm information are extracted to construct an alarm information time series, and the alarm information time series is input into a trained MOS value prediction model to obtain the MOS prediction value corresponding to the target cell identifier output by the MOS value prediction model, wherein the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model.

[0113] Further, such as Figure 4As shown, in terms of road test audit, after obtaining the MOS prediction value corresponding to the target cell identifier, the MOS prediction value corresponding to the target cell identifier is compared with a preset threshold (set to 2) to obtain a first comparison result; if the first comparison result indicates that the MOS prediction value corresponding to the target cell identifier is less than the preset threshold, a road test instruction and / or a network maintenance instruction is sent to the terminal device of the maintenance personnel corresponding to the target cell identifier; it can also be understood that if the MOS prediction value corresponding to the target cell identifier is identified as 1, a road test instruction and / or a network maintenance instruction is sent to the terminal device of the maintenance personnel corresponding to the target cell identifier.

[0114] Furthermore, if the maintenance personnel conducts an actual road test on the cell corresponding to the target cell identifier after receiving the road test instruction, the actual MOS value corresponding to the target cell identifier collected by the maintenance personnel is received, and it is determined whether the actual MOS value corresponding to the target cell identifier is consistent with the MOS predicted value corresponding to the target cell identifier. If the actual MOS value corresponding to the target cell identifier is inconsistent with the MOS predicted value corresponding to the target cell identifier, new sample data is generated based on the alarm record time series and the actual MOS value to fit and update the MOS value prediction model.

[0115] Further, such as Figure 4 As shown, in terms of customer care, in response to the first comparison result indicating that the MOS prediction value corresponding to the target cell identifier is less than a preset threshold, the model output time of the MOS prediction value corresponding to the target cell identifier is obtained; a target time period formed by tracing back a preset time length from the model output time is determined; at least one target user number related to the target cell identifier within the target time period is obtained; and a preset text is sent to the target user number.

[0116] Exemplarily, a specific implementation method of acquiring at least one target user number associated with a target cell identifier within a target time period and sending a preset text to the target user number is as follows:

[0117] For 2G users: Query all 2G-MC interface XDRs in the target time period, filter by CELL_ID, and obtain the user number.

[0118] For 4G users: Query all 4G-S1MME interface XDRs within the target time period, filter by CELL_ID, and obtain the user number.

[0119] For 5G users: Query all 5G-N1N2 interface XDRs in the target time period, filter by CELL_ID, and obtain the user number.

[0120] De-duplicate the retrieved user numbers, and retrieve the SMS sending records of each de-duplicated user number to ensure that no caring SMS has been sent within 24 hours; the user numbers that have not sent caring SMS within 24 hours after de-duplication are used as target user numbers, and the preset text (preset caring SMS) is sent to the target user numbers. This method helps to proactively care for users who have recently used cells with poor quality, thereby reducing complaints.

[0121] Further, such as Figure 4 As shown, in terms of geographic information system (GIS) map coloring, the previous MOS prediction value corresponding to the target cell identifier is obtained; the previous MOS prediction value corresponding to the target cell identifier is compared with the MOS prediction value corresponding to the target cell identifier at the current moment to obtain a second comparison result; in response to the second comparison result indicating that the previous MOS prediction value corresponding to the target cell identifier and the MOS prediction value corresponding to the target cell identifier at the current moment are different in value, a target color corresponding to the MOS prediction value corresponding to the target cell identifier at the current moment is obtained; and the color of the target position corresponding to the target cell identifier on the GIS map is updated based on the target color.

[0122] Among them, MOS=5 corresponds to dark green; MOS=4 corresponds to light green; MOS=3 corresponds to yellow; MOS=2 corresponds to orange; and MOS=1 corresponds to red.

[0123] Furthermore, the present application supports displaying the CELL_ID and MOS value by moving the mouse to the cell coverage area on the GIS map.

[0124] Generally speaking, monitoring personnel will focus on red areas. Red areas mean that the network service quality is poor. No red areas mean that the network operation quality is very good at this time; many red areas mean that the network operation quality is very poor at this time.

[0125] Figure 5 is a schematic diagram of a device for predicting MOS values ​​in a communication scenario shown in the present application, such as Figure 5 As shown, the prediction device 500 of the MOS value in the communication scenario includes a monitoring module 501, an acquisition module 502, a construction module 503 and a prediction module 504, wherein:

[0126] The monitoring module 501 is used to parse the real-time alarm information to extract the target cell identifier carried in the real-time alarm information after monitoring the real-time alarm information corresponding to the existing network element related to the mean opinion score MOS value.

[0127] The acquisition module 502 is used to acquire the latest N pieces of historical alarm information including real-time alarm information related to the target cell identifier, wherein the N pieces of historical alarm information are all alarm information related to the MOS value.

[0128] The construction module 503 is used to construct an alarm information time series based on the latest N historical alarm information.

[0129] The prediction module 504 is used to input the alarm information time series into a trained MOS value prediction model to obtain the MOS prediction value corresponding to the target cell identifier output by the MOS value prediction model, wherein the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model.

[0130] The device uses existing alarm information in the network to predict MOS values, eliminating the need for on-site measurements, thereby greatly reducing costs.

[0131] The prediction of the MOS value of this device is based on real-time alarm information, and the update is triggered by the alarm information, realizing the dynamic update of the MOS value. Compared with the traditional quarterly or monthly road test MOS collection method, this method provides higher real-time performance.

[0132] Traditional drive tests usually can only cover a limited area. This device uses alarm information to perform MOS prediction, which can cover all network cells and ensure overall monitoring of network performance.

[0133] Furthermore, the acquisition module 502 is also used to: extract historical alarm information related to the target cell identifier from the alarm information database; trace back the historical alarm information in chronological order starting from the real-time alarm information, and obtain the latest N historical alarm information including the real-time alarm information.

[0134] Furthermore, construction module 503 is also used to: extract key fields from each of the most recent N historical alarm information to obtain key field information corresponding to each of the most recent N historical alarm information; and construct an alarm information time series based on the key field information corresponding to each of the most recent N historical alarm information.

[0135] Furthermore, the device 500 for predicting MOS values ​​in a communication scenario also includes a training module, which is used to: according to an embodiment of the present application, obtain multiple groups of drive test data, wherein each group of drive test data includes 1 drive test cell identifier, 1 drive test collection time and 1 drive test MOS actual value; for any group of drive test data, obtain historical alarm information samples corresponding to the drive test cell identifier included in the group of drive test data, and trace back the historical alarm information samples from the drive test collection time included in the group of drive test data to obtain N traced historical alarm information samples; for any group of drive test data, construct training samples according to the group of drive test data and the N historical alarm information samples corresponding to the group of drive test data; obtain training samples corresponding to each of the multiple groups of drive test data to form a training sample set, and train the initial MOS value prediction model based on the training sample set until the training is completed, and obtain the MOS value prediction model generated by the training.

[0136] Furthermore, the training module is also used to: for any group of road test data, construct an alarm information time series sample based on N historical alarm information samples corresponding to the group of road test data, and use the actual road test MOS value included in the group of road test data as the sample label of the alarm information time series sample; for any group of road test data, form a training sample based on the alarm information time series sample and sample label corresponding to the group of road test data.

[0137] Furthermore, the training module is also used to: for any group of drive test data, jointly construct an alarm information time series sample based on the road test cell identifier, the road test collection time and N historical alarm information samples corresponding to the group of road test data, and use the road test cell identifier and the actual value of the road test MOS included in the group of road test data as sample labels of the alarm information time series sample; for any group of drive test data, form a training sample based on the alarm information time series samples and sample labels corresponding to the group of road test data.

[0138] Furthermore, the device 500 for predicting MOS values ​​in a communication scenario also includes a post-processing module, which is used to: compare the MOS prediction value corresponding to the target cell identifier with a preset threshold to obtain a first comparison result; in response to the first comparison result indicating that the MOS prediction value corresponding to the target cell identifier is less than the preset threshold, send a road test instruction and / or a network maintenance instruction to the terminal device of the maintenance personnel corresponding to the target cell identifier.

[0139] Furthermore, the post-processing module is also used to: in response to the first comparison result indicating that the MOS prediction value corresponding to the target cell identifier is less than a preset threshold, obtain the model output time of the MOS prediction value corresponding to the target cell identifier; determine a target time period formed by tracing back a preset time length from the model output time; obtain at least one target user number related to the target cell identifier within the target time period; and send a preset text to the target user number.

[0140] Furthermore, the post-processing module is also used to: obtain the previous MOS prediction value corresponding to the target cell identifier; compare the previous MOS prediction value corresponding to the target cell identifier with the MOS prediction value corresponding to the target cell identifier at the current moment to obtain a second comparison result; in response to the second comparison result indicating that the previous MOS prediction value corresponding to the target cell identifier and the MOS prediction value corresponding to the target cell identifier at the current moment are different in value, obtain the target color corresponding to the MOS prediction value corresponding to the target cell identifier at the current moment; and update the color of the target position corresponding to the target cell identifier on the geographic information system GIS map based on the target color.

[0141] In order to implement the above embodiment, the present application embodiment further proposes an electronic device 600, such as Figure 6 As shown, the electronic device 600 includes: a processor 601 and a memory 602 communicatively connected to the processor, the memory 602 stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor 601 to implement the method for predicting MOS values ​​in the communication scenario as shown in the above embodiment.

[0142] In order to implement the above embodiment, the embodiment of the present application also proposes a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to implement the method for predicting MOS values ​​in a communication scenario as shown in the above embodiment.

[0143] In order to implement the above embodiment, the embodiment of the present application also proposes a computer program product, including a computer program, which, when executed by a processor, implements the method for predicting the MOS value in the communication scenario as shown in the above embodiment.

[0144] In the description of the present application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0145] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0146] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0147] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0148] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for predicting MOS value in a communication scenario, characterized in that: include: After monitoring the real-time alarm information corresponding to the existing network element related to the mean opinion score MOS value, parsing the real-time alarm information to extract the target cell identifier carried by the real-time alarm information; Acquire the latest N historical alarm information related to the target cell identifier, including the real-time alarm information, wherein the N historical alarm information are all alarm information related to the MOS value; Constructing an alarm information time series based on the most recent N pieces of historical alarm information; The alarm information time series is input into a trained MOS value prediction model to obtain a MOS prediction value corresponding to the target cell identifier output by the MOS value prediction model, wherein the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model.

2. The method according to claim 1, characterized in that The acquiring of the most recent N pieces of historical alarm information including the real-time alarm information related to the target cell identifier includes: Extracting historical alarm information related to the target cell identifier from an alarm information database; The historical alarm information is traced back in chronological order starting from the real-time alarm information to obtain the latest N historical alarm information including the real-time alarm information.

3. The method according to claim 2, characterized in that The constructing the alarm information time series based on the most recent N pieces of historical alarm information includes: Extract key fields from each of the most recent N pieces of historical alarm information to obtain key field information corresponding to each of the most recent N pieces of historical alarm information; The alarm information time series is constructed based on the key field information corresponding to each of the most recent N pieces of historical alarm information.

4. The method according to any one of claims 1 to 3, characterized in that The training method of the MOS value prediction model comprises: Acquire multiple groups of drive test data, wherein each group of the drive test data includes a drive test cell identifier, a drive test collection time, and a drive test MOS actual value; For any group of the drive test data, obtain historical alarm information samples corresponding to the drive test cell identifiers included in the group of drive test data, and trace back the historical alarm information samples from the drive test collection time included in the group of drive test data to obtain N traced historical alarm information samples; For any group of the drive test data, construct a training sample according to the group of drive test data and N historical alarm information samples corresponding to the group of drive test data; The training samples corresponding to the plurality of groups of drive test data are obtained to form a training sample set, and the initial MOS value prediction model is trained based on the training sample set until the training is completed, and the MOS value prediction model generated by the training is obtained.

5. The method according to claim 4, characterized in that The step of constructing a training sample for any group of the drive test data according to the group of the drive test data and N historical alarm information samples corresponding to the group of the drive test data includes: For any group of the drive test data, construct an alarm information time series sample according to the N historical alarm information samples corresponding to the group of drive test data, and use the actual drive test MOS value included in the group of drive test data as the sample label of the alarm information time series sample; For any group of the drive test data, training samples are formed based on the warning information time series samples and the sample labels corresponding to the group of drive test data.

6. The method according to claim 4, characterized in that The step of constructing a training sample for any group of the drive test data according to the group of the drive test data and N historical alarm information samples corresponding to the group of the drive test data includes: For any group of the drive test data, construct an alarm information time series sample according to the drive test cell identifier, the drive test collection time and N historical alarm information samples corresponding to the group of drive test data, and use the drive test cell identifier and the actual drive test MOS value included in the group of drive test data as sample labels of the alarm information time series sample; For any group of the drive test data, training samples are formed based on the warning information time series samples and the sample labels corresponding to the group of drive test data.

7. The method according to any one of claims 1 to 3, characterized in that The method further comprises: Comparing the MOS prediction value corresponding to the target cell identifier with a preset threshold to obtain a first comparison result; In response to the first comparison result indicating that the MOS prediction value corresponding to the target cell identifier is less than the preset threshold, a drive test instruction and / or a network maintenance instruction is sent to a terminal device of a maintenance personnel corresponding to the target cell identifier.

8. The method according to claim 7, characterized in that The method further comprises: In response to the first comparison result indicating that the MOS prediction value corresponding to the target cell identifier is less than the preset threshold, obtaining a model output time of the MOS prediction value corresponding to the target cell identifier; Determine a target time period formed by tracing back a preset time period from the time when the model is output; Acquire at least one target user number associated with the target cell identifier within the target time period; Send a preset text to the target user number.

9. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: Obtaining a previous MOS prediction value corresponding to the target cell identifier; Compare the last MOS prediction value corresponding to the target cell identifier with the MOS prediction value corresponding to the target cell identifier at the current moment to obtain a second comparison result; In response to the second comparison result indicating that the values ​​of the previous MOS prediction value corresponding to the target cell identifier and the MOS prediction value corresponding to the target cell identifier at the current moment are different, obtaining a target color corresponding to the MOS prediction value corresponding to the target cell identifier at the current moment; The color of the target position corresponding to the target cell identifier on a geographic information system (GIS) map is updated based on the target color.

10. A device for predicting MOS value in a communication scenario, characterized in that: include: A monitoring module, configured to parse the real-time alarm information corresponding to the existing network element related to the mean opinion score MOS value after monitoring the real-time alarm information to extract the target cell identifier carried by the real-time alarm information; An acquisition module, used to acquire the latest N historical alarm information related to the target cell identifier including the real-time alarm information, wherein the N historical alarm information are all alarm information related to the MOS value; A construction module, used to construct an alarm information time series based on the latest N historical alarm information; The prediction module is used to input the alarm information time series into a trained MOS value prediction model to obtain the MOS prediction value corresponding to the target cell identifier output by the MOS value prediction model, wherein the MOS value prediction model is a voice MOS value prediction model or a video MOS value prediction model.

11. An electronic device, comprising: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 9.

12. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-9.

13. A computer program product comprising a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 9.

Images