Link risk prediction method and device, electronic equipment and computer readable medium

Abstract

The invention discloses a link risk prediction method and device, electronic equipment and a computer readable medium. The method comprises the following steps: acquiring a target service index related to link underlying characteristics in a target network; determining a link degradation threshold based on the target service index and a preset service index degradation threshold; and predicting the risk of the link in the target network based on the link degradation threshold and the link data of the target network within the preset time period to obtain a link risk prediction result. According to the technical scheme provided by the invention, the accuracy and credibility of link degradation prediction can be improved.

Description

Technical Field

[0001] The present application relates to communication technology, and in particular to a link risk prediction method and device, electronic equipment, and computer-readable medium. Background Art

[0002] Due to the huge number of optical fiber links in base stations, the high construction risks, and the numerous data alarms in the field, the maintenance of the link network is a difficult point in base station operation and maintenance. Although the risk health of the current network can be identified through big data and artificial intelligence (AI) algorithms, the big data analysis technology currently used is based on the analysis of the key performance indicators (KPI) data and alarm data of the optical link, which directly gives the judgment of the link health level, breaking away from the limitations of business indicators. Not only is the credibility of the judgment result not enough to satisfy users, but this judgment result cannot accurately predict the risk of link degradation. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to propose a link risk prediction method and device, an electronic device, and a computer-readable medium, which can improve the accuracy and credibility of link degradation prediction.

[0004] In a first aspect, an embodiment of the present application provides a link risk prediction method, comprising:

[0005] Obtain target service indicators related to the underlying characteristics of the link in the target network;

[0006] Determining a link degradation threshold based on the target service indicator and a preset service indicator degradation threshold;

[0007] The risk of the link in the target network is predicted based on the link degradation threshold and the link data of the target network within a preset time period to obtain a link risk prediction result.

[0008] In a second aspect, an embodiment of the present application provides a link risk prediction device, including:

[0009] An acquisition module, used to acquire target service indicators related to the underlying characteristics of the link in the target network;

[0010] A determination module, configured to determine a link degradation threshold based on the target service indicator and a preset service indicator degradation threshold;

[0011] The prediction module is used to predict the risk of the link in the target network based on the link degradation threshold and the link data of the target network within a preset time period to obtain a link risk prediction result.

[0012] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0013] one or more processors;

[0014] A storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the link risk prediction method provided in the embodiment of the present application;

[0015] One or more I / O interfaces are connected between the processor and the memory and are configured to implement information interaction between the processor and the memory.

[0016] In a fourth aspect, an embodiment of the present application provides a storage medium on which a computer program is stored, and when the program is executed by a processor, the method for link risk prediction provided by the embodiment of the present application is implemented.

[0017] The link risk prediction method provided in the embodiment of the present application obtains the target business indicator related to the underlying characteristics of the link in the target network, and then determines the link degradation threshold based on the target business indicator and the preset business indicator degradation threshold, so that the upper-layer target business indicator is dynamically associated with the link degradation threshold of the underlying link, and then predicts the risk of the link in the target network based on the link degradation threshold and the link data of the target network within a preset time period to obtain a link risk prediction result, so that the basis for link risk prediction is more comprehensive, not only related to the link data at the link level, but also related to the business indicators at the business level. It can not only accurately predict the future risks of the link, such as issuing alarm information in advance, so that the operation and maintenance personnel can eliminate the risks of each link in the target network in advance, avoid the losses caused by blind operation of the operation and maintenance personnel, but also improve user satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of a link risk prediction method is provided for an embodiment of the present application;

[0019] Figure 2 A schematic diagram of the structure of a risk prediction model is provided for an embodiment of the present application;

[0020] Figure 3 A flowchart for determining a link degradation threshold provided in an embodiment of the present application;

[0021] Figure 4 This is a functional block diagram of a link risk prediction device according to an embodiment of the present application;

[0022] Figure 5 A principle block diagram of an electronic device is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the technical solution of the present application, the server provided by the present application is described in detail below in conjunction with the accompanying drawings.

[0024] Example embodiments will be described more fully below with reference to the accompanying drawings, but the example embodiments may be embodied in different forms and should not be construed as limited to the embodiments set forth herein. On the contrary, the purpose of providing these embodiments is to make this application thorough and complete and to enable those skilled in the art to fully understand the scope of this application.

[0025] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0026] The terms used herein are only used to describe specific embodiments and are not intended to limit the present application. As used herein, the singular forms "a", "an" and "the" are also intended to include the plural forms, unless the context clearly indicates otherwise. It will also be understood that when the terms "comprising" and / or "made of" are used in this specification, the presence of the features, wholes, steps, operations, elements and / or components is specified, but the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or groups thereof is not excluded.

[0027] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this application, and will not be interpreted as having an idealized or overly formal meaning unless explicitly defined herein.

[0028] The number of base station fiber links is huge and the links are complex. In addition, the link quality of different sites varies. If the link degradation threshold is used to predict link risk, the prediction results will be inaccurate. The current artificial intelligence analysis predicts link risk based on link data, without considering the impact of business indicators on link operation status. Therefore, the credibility of the prediction results is low, resulting in low user satisfaction with the prediction results. In addition, due to inaccurate prediction results, it is easy for operation and maintenance personnel to operate blindly, causing unnecessary losses to network operation.

[0029] In a first aspect, an embodiment of the present application provides a link risk prediction method, which can improve the accuracy and feasibility of link risk prediction.

[0030] Figure 1 A flowchart of a link risk prediction method is provided for an embodiment of the present application. Figure 1 , the embodiment of the present application provides a link risk prediction method, comprising:

[0031] Step S101, obtaining target service indicators related to link bottom layer characteristics in a target network.

[0032] The embodiment of the present application does not limit the target network, and the target network may be a wireless communication network. For example, the target network may be a fiber optic network where a bureau is located, and the fiber optic network corresponding to each bureau includes multiple cells.

[0033] The link bottom layer characteristics refer to the characteristics that affect the packet error rate of the target network, and specifically may be the characteristics that affect the quality of the optical link of the wireless communication network.

[0034] The target service indicator is a service indicator related to the underlying characteristics of the link. Service indicators include, but are not limited to, Radio Resource Control (RRC) establishment success rate, User Equipment (UE) context establishment success rate, Protocol Data Unit Session (PDUSession) establishment success rate, Intra-system handover success rate, Intra-system handover success rate, Wireless drop rate, Wireless connection rate. The target service indicator can be one or more of the service indicators.

[0035] In some embodiments, obtaining a target service indicator related to a link bottom layer characteristic in a target network includes:

[0036] Based on the mean value of each candidate business indicator in multiple valid intervals and the interval value of the valid interval, multiple fitting points are obtained, and the fitting points corresponding to each candidate business indicator are fitted using a fitting method to obtain a fitting result corresponding to each candidate business indicator; based on the fitting result, the association parameters between the candidate business indicator and the packet error rate are determined; based on the association parameters between the candidate business indicator and the packet error rate and a preset parameter threshold, the target business indicator is determined.

[0037] The valid interval is the packet error interval when the number of packet errors in the packet error interval is greater than a preset packet error threshold. The packet error interval is an interval determined based on the optical port packet error rate. The optical port packet error rate refers to the packet error rate of the optical communication interface used to transmit link data. The fitting method includes a filter fitting method, and specifically, a convolution smoothing (Savitzky-Golay) filter fitting method can be used.

[0038] In some embodiments, the correlation parameters include a significance level and a Pearson correlation coefficient, and correspondingly, the parameter thresholds include a significance threshold and a Pearson threshold.

[0039] In some embodiments, the target business indicator is determined based on the association parameters between the candidate business indicator and the packet error rate and a preset parameter threshold, including: when the significance level is less than the significance threshold and the Pearson correlation coefficient is greater than the Pearson threshold, the candidate business indicator is determined as the target business indicator.

[0040] The significance level and Pearson correlation coefficient between each candidate service indicator and the packet error rate; based on the significance level, the Pearson correlation coefficient, and the preset significance threshold and Pearson threshold, determine the target service indicator.

[0041] For example, the packet error rate of 0-0.1% is divided into multiple packet error intervals, and the interval width of each interval can be set to 0.005%, thereby obtaining 20 packet error intervals. The number of packet errors in each packet error interval is counted, and when the number of packet errors is greater than the preset packet error valid value of 100, the packet error interval is a valid interval.

[0042] It should be noted that the present application embodiment does not limit the interval width of the error packet interval, the number of error packet intervals and the error packet effective value. The interval width of the error packet interval and the number of error packet intervals can be arbitrarily set by the user according to the actual situation of the target network, and the error packet effective value can also be set by the user.

[0043] In the embodiment of the present application, within any valid interval, the mean value of each candidate service indicator and the interval value of the valid interval can determine a fitting point. For example, for the RRC establishment success rate, within any valid interval, the mean value of the RRC establishment success rate is used as the ordinate of the fitting point, and the interval value of the valid interval is used as the abscissa to determine a fitting point.

[0044] The embodiment of the present application does not limit the interval value of the valid interval. For example, the interval value of the valid interval may be the middle value of the valid interval or a specified value within the valid interval.

[0045] In some embodiments, for any candidate business indicator, the Savitzky-Golay filter fitting method is used to fit the fitting points corresponding to the candidate business indicator to obtain the fitting result corresponding to the candidate business indicator, that is, each candidate business indicator can obtain a fitting result.

[0046] The significance level and the Pearson correlation coefficient between each candidate service indicator and the packet error rate are determined based on the fitting result, wherein the significance level and the Pearson correlation coefficient can be determined based on the fitting result by using the currently available method.

[0047] The embodiment of the present application does not limit the significance threshold and the Pearson threshold. For example, the significance threshold may be 0.001, and the Pearson threshold may include a positive correlation strength threshold and a negative correlation strength threshold. The positive correlation strength threshold may be 0.9, and the negative correlation strength threshold may be -0.9. When the significance level is less than 0.001, and the positive correlation strength threshold is greater than 0.9 or the negative correlation strength threshold is less than -0.9, it is determined that the candidate service indicator is strongly correlated with the underlying link characteristics, and the candidate service indicator is determined as the target service indicator. The target service indicator can be used as a service indicator for generating a link degradation threshold.

[0048] Step S102: determining a link degradation threshold based on the target service indicator and a preset service indicator degradation threshold.

[0049] The embodiment of the present application does not limit the method for obtaining the service indicator degradation threshold. For example, the service indicator degradation threshold can be determined based on experience or based on statistical analysis.

[0050] In some embodiments, a link degradation threshold is determined based on a target business indicator and a preset business indicator degradation threshold, including: obtaining a link packet error threshold corresponding to each target business indicator based on a fitting result; and taking the minimum value of the link packet error thresholds corresponding to the business indicator degradation threshold as the link degradation threshold.

[0051] For any target service indicator, fitting is performed based on the link packet error data and the target service data to obtain the fitting result corresponding to the target service indicator. The link packet error threshold can be obtained according to the fitting result, that is, each target service indicator corresponds to a link packet error threshold, and the minimum value of the link packet error thresholds corresponding to these target service indicators is used as the link degradation threshold.

[0052] In some embodiments, the link error packet threshold includes an uplink error packet threshold and a downlink error packet threshold. When determining the link degradation threshold, the lowest value of the uplink error packet threshold and the downlink error packet threshold corresponding to the service indicator degradation threshold can be used as the link degradation threshold of the target network.

[0053] Step S103 , predicting the risk of links in the target network based on the link degradation threshold and the link data of the target network within a preset time period, and obtaining a link risk prediction result.

[0054] The preset time period may be a certain time period before the current moment, for example, taking the current moment as the starting point and moving forward 7 days as the preset time period; or taking the current moment as the starting point and moving forward 30 days as the preset time period.

[0055] After determining the link degradation threshold, the risk of the link in the target network is predicted based on the link data of the target network in the preset time period before the current moment. For example, the risk of the link in the target network is predicted based on the link data of the target network within 60 days before the current moment to obtain a link risk prediction result. It should be noted that 60 days is an exemplary description and is not a limitation on the preset time period.

[0056] In some embodiments, the link data includes link error packet data and alarm data. The link error packet data and alarm data within 60 days before the current moment are used to predict the risk of the link in the target network to obtain a link risk prediction result.

[0057] The disclosed embodiment predicts the risk of links in a target network through a risk prediction model based on a link degradation threshold and link data of the target network within a preset time period.

[0058] Specifically, the link degradation threshold and the link data of the target network in a preset time period before the current moment are input into the risk prediction model, the risk of the link in the target network is predicted, and the link risk prediction result is obtained.

[0059] In some embodiments, the risk prediction model is obtained by training through the following steps S21-S23:

[0060] Step S21, standardize the link error packet data to obtain a standard link error packet sample.

[0061] The link degradation threshold includes an uplink degradation threshold and a downlink degradation threshold, and the link error packet data includes uplink error packet data and downlink error packet data.

[0062] In some embodiments, link error packet data samples are standardized to obtain standard link error packet data, including: using a packet error rate standardization processing function to process uplink error packet data in the link error packet data based on an uplink degradation threshold to obtain standard uplink error packet data, and processing downlink error packet data in the link error packet data based on a downlink degradation threshold to obtain standard downlink error packet data.

[0063] In some embodiments, the packet error rate normalization processing function is used to amplify the relationship between the collected packet error data and the link degradation threshold. For example, the relationship between the packet error data and the link degradation threshold can be a numerical value, that is, the relationship between the packet error data and the link degradation threshold is represented by a numerical value, and the packet error rate normalization processing function is a function that can amplify the numerical value.

[0064] The packet error rate normalization processing function in the embodiment of the present application may be an improved sigmoid activation function. Specifically, the improved sigmoid activation function is:

[0065]

[0066] Among them, i represents the time series, x i represents the uplink error packet data or downlink error packet data at the i-th moment, y i represents standard uplink error packet data or standard downlink error packet data, and T represents the link degradation threshold (uplink degradation threshold or downlink degradation threshold).

[0067] Step S22, standardize the alarm data to obtain a standard alarm sample.

[0068] In some embodiments, the alarm data is standardized to obtain standard alarm data, including: obtaining an alarm time period of each alarm based on the alarm data; and mapping the alarm time period to a preset alarm time point to obtain standard alarm data.

[0069] For any alarm data, first obtain the alarm time period of the alarm, which includes the alarm time and alarm duration, and then map the alarm time period to the alarm time point. For example, if an alarm occurs at a certain time point, it is marked as "1"; if no alarm occurs at a certain time point, it is marked as "0". When the alarm duration is long and spans multiple time points, multiple time points can be marked as "1".

[0070] For example, if the alarm data includes an RRU link disconnection alarm and an optical port link failure alarm, the alarm time period of the RRU link disconnection alarm is obtained. If the alarm time period of the RRU link disconnection alarm includes an alarm time of 1:00 and an alarm duration of 5 minutes, if the alarm time point is 1:00, then the duration of the RRU link disconnection alarm is mapped to 1:00. If the alarm time period of the RRU link disconnection alarm includes an alarm time of 1:00 and an alarm duration of 1 hour, if the alarm time points include 1:00 and 1:30, then the duration of the RRU link disconnection alarm is mapped to 1:00 and 1:30, that is, the alarm time points 1:00 and 1:30 are both marked as "1".

[0071] Step S23, inputting the risk prediction model for training based on the standard link error packet samples and the standard alarm samples, and obtaining the trained risk prediction model after the risk prediction model converges.

[0072] In some embodiments, Figure 2As shown in FIG. 1 , the structure of the risk prediction model adopts a two-layer long short-term memory network (LSTM) structure. Each layer output is connected to the Dropout layer to randomly discard the neural unit, and finally the Dense layer is used to output the alarm prediction result. Specifically, the risk prediction model includes a first LSTM layer 21, a second LSTM layer 23, a first Dropout layer 22, a second Dropout layer 24 and a fully connected layer (Dense layer) 25. The first LSTM layer receives the error packet data and the alarm data X 1 , X 2 , X 3 ...X n The data output by the first LSTM layer 21 is input into the first Dropout layer 22, the data output by the first Dropout layer 22 is input into the second LSTM layer 23, the data output by the second LSTM layer 23 is input into the second Dropout layer 24, the second Dropout layer 24 is input into the fully connected layer 25, and the fully connected layer 25 outputs the predicted link risk prediction result y 1 ,y 2 ,y 3 ……y n .

[0073] In the embodiment of the present application, the first Dropout layer 21 and the second Dropout layer 23 randomly discard neural units to avoid overfitting, and finally the fully connected layer outputs the predicted link risk prediction result.

[0074] In some embodiments, the standard link error packet samples and standard alarm samples for training the risk prediction model are input into the risk prediction model to be trained in units of links in the following format. Each link includes four features, namely, two alarm data, uplink error packet data, and downlink error packet data, and 60 days are divided into 5760 time points.

[0075] In some embodiments, the standard link error packet samples and the standard alarm samples are divided into a training set, a test set and a validation set, and the ratio of the number of samples in the training set, the test set and the validation set can be 8:1:1.

[0076] During the model training process, the cross entropy is calculated as the loss function, and the gradient descent is used to iterate repeatedly until the risk prediction model converges to obtain a trained risk prediction model.

[0077] In an embodiment of the present application, the trained risk prediction model is used to predict the link risk within a period of time after the current moment. For example, the link risk 7 days after the current moment can be predicted, that is, the link risk prediction result 7 days after the current moment is obtained.

[0078] In some embodiments, the risk of links in the target network is predicted based on the link degradation threshold and the link data of the target network in a preset time period before the current moment. After obtaining the link risk prediction result, it also includes: determining the risk level of the link based on the link risk prediction result.

[0079] The link risk prediction result includes the number of alarms, and the risk level of the link can be divided according to the predicted number of alarms. Table 1 shows the relationship between the number of alarms and the risk level.

[0080] Table 1

[0081]

[0082] The link risk prediction result also includes information such as link identification, time period in which the alarm will occur, etc. For example, link A will generate an alarm from 14:00 to 14:15 on the next day of the current time.

[0083] In some embodiments, before obtaining the target service indicator related to the link bottom layer characteristics in the target network, the method further includes: preliminarily screening the service data corresponding to each service indicator, retaining only one group of service data with the same change trend within a preset time period as candidate service data; and cleaning the candidate service data to obtain valid service data. The cleaning includes one or more of service interference cleaning and noise cleaning.

[0084] The collection module is used to collect the target network's full-network link packet error data, cell service data, and alarm data. The collection time length can be set by the user. For example, the link packet error data, cell service data, and alarm data within 60 days before the current time can be collected. The cell service data collection granularity can be 15 minutes, that is, the cell service indicator data is collected every 15 minutes.

[0085] During the initial screening, the business data is screened according to different business indicators.

[0086] For example, a dynamic time warping algorithm is used to preliminarily screen all service data, and only one service data with high trend change similarity within a preset time period (such as 30 days or 60 days, etc.) is retained as candidate service data for generating link degradation thresholds. For candidate service data, the link and cell where the candidate service data is located are obtained by querying the pre-configured configuration information, and the corresponding relationship between the link and the cell is established, that is, sample data with one-to-one association between link error packet data and cell service data is generated.

[0087] In some embodiments, service interference cleaning includes one or more of the following methods:

[0088] In some embodiments, service interference cleaning includes determining links where no packet errors have occurred from service data, and removing service data corresponding to the links where no packet errors have occurred. Specifically, links where the packet error rate of uplink service data and downlink service data at all time granularity points is 0 are regarded as invalid service data, and all service data related to the link are removed, that is, the service data of the invalid link is removed.

[0089] In some service data, there may be interference between uplink service data and downlink service data. In order to improve the accuracy of the link degradation threshold, service interference cleaning may also include determining service data with interference between uplink and downlink from the service data and removing it. Specifically, service data at the time point where both uplink and downlink error packets exist are removed based on the service data, that is, service data with interference between uplink and downlink is removed.

[0090] Service interference cleaning can also include determining and removing service data affected by environmental interference from service data, that is, removing time-granularity point data at high interference moments according to service indicators. For example, if the continuous quality improvement (CQI) excellent rate is less than 80%, it means that the downlink wireless environment is seriously interfered, and the uplink average interference level is greater than -105dbm, it means that the uplink wireless environment is seriously interfered. The service data at the time granularity point that does not meet the service indicators is removed.

[0091] According to the alarm data, the business data of the time granularity points with other interference factors are eliminated, that is, the alarm time period of each alarm is obtained based on the alarm data; the alarm time period is mapped to the preset alarm time point to obtain standard alarm data.

[0092] In some embodiments, the service data may also be subjected to denoising, that is, the service data interfered by noise may be removed. Specifically, the link error packet data may be partitioned to obtain multiple error packet intervals, each of which may have a width of 0.0005%, and the k-means clustering algorithm may be used to remove outliers in the service data.

[0093] In some embodiments, after initially screening the service data corresponding to each service indicator and retaining only one group of service data with the same change trend within a preset time period as candidate service data, the process further includes: determining the link where the candidate service data is located, and obtaining the corresponding relationship between the link where the candidate service data is located and the cell. When the level of the risk prediction result of the link is high, the operation and maintenance personnel can quickly determine the cell to quickly eliminate the target network risk.

[0094] Figure 3 A flowchart for determining a link degradation threshold is provided in an embodiment of the present application, such as Figure 3 As shown, the link degradation threshold is determined by the following steps:

[0095] Step S301, obtaining the error packet data of the whole network link, the service data of the cell and the alarm data within a preset time period.

[0096] The closer the whole network link packet error data, cell service data and alarm data are to the current time, the closer they are to the actual operation status of the link. Therefore, the whole network link packet error data, cell service data and alarm data within 30 days before the current time are obtained.

[0097] Step S302: initial screening of business indicator similarity.

[0098] A dynamic time warping algorithm is used to obtain candidate business data by retaining only one business data for business indicators with high similarity in trend changes within 30 days.

[0099] Step S303: pairing link error packet data with candidate service data.

[0100] For the selected service indicator, the link error packet data is paired with the candidate service data by querying the pre-configured cell configuration information, that is, a one-to-one association between the link error packet data and the candidate service data is obtained.

[0101] Step S304: clean the candidate business data and alarm data.

[0102] Cleaning of candidate business data includes:

[0103] The links with both uplink error packet data and downlink error packet data of 0 at all time granularity points within 30 days are eliminated, that is, the candidate service data corresponding to the links without error packets are eliminated to reduce the subsequent data processing volume.

[0104] The service data with interference between uplink and downlink is eliminated, that is, the service data at the time point where both uplink and downlink error packets exist are eliminated based on the service data, that is, the service data with interference between uplink and downlink is eliminated.

[0105] Eliminate the time-granularity business data of high-interference moments according to business indicators. For example, a CQI quality rate lower than 80% indicates that the downlink wireless environment is severely interfered, and an uplink average interference level greater than -105dbm indicates that the uplink wireless environment is severely interfered. All time-granularity data that do not meet business indicators must be eliminated.

[0106] Cleaning of alarm data includes:

[0107] Eliminate business data at time granularity points where other interference factors exist based on the alarm. For example, map the duration of the alarm to the corresponding time point. The occurrence of the alarm is represented by 1, otherwise it is 0. Eliminate business data at time granularity points where the alarm has occurred.

[0108] Based on the cleaned business data, valid business data is obtained, and the valid business data is binned and denoised. The link error packet data is divided into intervals (interval width is 0.005%) to obtain multiple error packet intervals. The k-means clustering algorithm is used to remove outliers in the link error packet data. The number of error packets in each error packet interval is counted. If it is greater than 100, the error packet interval is determined as a valid interval. The link error packet data and business data in the valid interval can be used to determine the link error packet threshold.

[0109] Step S305: screening the correlation between link error packet data and service indicators to determine target service indicators.

[0110] Based on the mean value of each candidate business indicator in multiple valid intervals and the interval value of the valid interval, multiple fitting points are obtained, and the fitting points corresponding to each candidate business indicator are fitted using the Savitzky-Golay filter fitting method to obtain the fitting results corresponding to each candidate business indicator. Based on the fitting results, the significance level and Pearson correlation coefficient between each candidate business indicator and the packet error rate are determined, and based on the significance level, the Pearson correlation coefficient, and the preset significance threshold and Pearson threshold, the target business indicator is determined.

[0111] For example, the significance threshold may be 0.001, the Pearson threshold may include a positive correlation strength threshold and a negative correlation strength threshold, the positive correlation strength threshold may be 0.9, and the negative correlation strength threshold may be -0.9. When the significance level is less than 0.001, and the positive correlation strength threshold is greater than 0.9 or the negative correlation strength threshold is less than -0.9, it is determined that the candidate service indicator is strongly correlated with the underlying link characteristic, and the candidate service indicator is determined as the target service indicator.

[0112] Step S306: determining a link degradation threshold based on the target service indicator and a preset service indicator degradation threshold.

[0113] Based on the fitting result, the link packet error threshold corresponding to the target service indicator in each valid interval is obtained; and the minimum value of the link packet error thresholds corresponding to the service indicator degradation threshold is used as the link degradation threshold.

[0114] After determining the link degradation threshold, the link error packet data and alarm data within a preset time period can be used to train the risk prediction model, and the trained risk prediction model can be used to predict the alarm information within a period of time in the future (such as 7 days). The alarm information includes but is not limited to the number of alarms, the link where the alarm occurs, and other information. The link risk level can also be determined based on the number of alarms, which helps operation and maintenance personnel take corresponding measures based on the link risk level.

[0115] In some embodiments, the link degradation threshold can be used not only to predict link risks, but also to detect link risks. For example, the packet error rate of the current link is obtained. If the packet error rate of the current link is greater than or equal to the link degradation threshold, the current link is judged to be a risky link. If the packet error rate of the previous link is less than the link degradation threshold, the current link is judged to be risk-free. The link risk prediction method provided in the embodiment of the present application obtains the target business indicator related to the underlying characteristics of the link in the target network, and then determines the link degradation threshold based on the target business indicator and the preset business indicator degradation threshold, so that the upper target business indicator is dynamically associated with the link degradation threshold of the link bottom layer, and then predicts the risk of the link in the target network based on the link degradation threshold and the link data of the target network in the preset time period, and obtains the link risk prediction result, so that the basis for link risk prediction is more comprehensive, not only related to the link data at the link level, but also related to the business indicators at the business level, and can not only accurately predict the future risks of the link, such as issuing alarm information in advance, so that the operation and maintenance personnel can eliminate the risks of each link in the target network in advance, avoid the losses caused by blind operation of the operation and maintenance personnel, but also improve user satisfaction.

[0116] In a second aspect, an embodiment of the present application provides a link risk prediction device. Figure 4 Schematic diagram of the link risk prediction device of the embodiment of the present application. Figure 4 The link risk prediction device 400 provided in the embodiment of the present application includes:

[0117] An acquisition module 401 is used to acquire target service indicators related to the underlying characteristics of the link in the target network;

[0118] A determination module 402, configured to determine a link degradation threshold based on a target service indicator and a preset service indicator degradation threshold;

[0119] The prediction module 403 is used to predict the risk of the link in the target network based on the link degradation threshold and the link data of the target network in a preset time period to obtain a link risk prediction result.

[0120] In some embodiments, the acquisition module 401 is also used to obtain multiple fitting points based on the mean of each candidate service indicator in multiple valid intervals and the interval value of the valid interval, the valid interval is a packet error interval that meets the preset conditions, and the packet error interval is an interval determined based on the optical port packet error rate; the fitting point corresponding to each candidate service indicator is fitted by the fitting method to obtain the fitting result corresponding to each candidate service indicator; the association parameter between the candidate service indicator and the packet error rate is determined based on the fitting result, and then the target service indicator is determined based on the association parameter between the candidate service indicator and the packet error rate and the preset parameter threshold. Among them, the association parameter includes the significance level and the Pearson correlation coefficient, and correspondingly, the parameter threshold includes the significance threshold and the Pearson threshold. Based on the significance level, the Pearson correlation coefficient, and the preset significance threshold and the Pearson threshold, the target service indicator is determined.

[0121] In some embodiments, the determination module 402 is further used to obtain a link packet error threshold corresponding to each target service indicator based on the fitting result; and use the minimum value of the link packet error thresholds corresponding to the service indicator degradation thresholds as the link degradation threshold.

[0122] In some embodiments, the preset condition includes: when the number of error packets in the error packet interval is greater than a preset error packet valid value, the error packet interval is a valid interval.

[0123] In some embodiments, the link data includes link error packet data and alarm data.

[0124] In some embodiments, the prediction module 403 is further used to input the link degradation threshold and the link data of the target network within a preset time period into the risk prediction model, predict the risk of the link in the target network, and obtain the link risk prediction result.

[0125] The risk prediction model is trained through the following steps: standardizing the link error packet data to obtain standard link error packet samples; standardizing the alarm data to obtain standard alarm samples; training the risk prediction model based on the standard link error packet samples and standard alarm samples, and obtaining a trained risk prediction model after the risk prediction model converges.

[0126] In some embodiments, the link degradation threshold includes an uplink degradation threshold and a downlink degradation threshold, and the link error packet data includes uplink error packet data and downlink error packet data;

[0127] The link error packet data samples are standardized to obtain standard link error packet data, including: using a packet error rate standardization processing function to process uplink error packet data in the link error packet data based on an uplink degradation threshold to obtain standard uplink error packet data, and processing downlink error packet data in the link error packet data based on a downlink degradation threshold to obtain standard downlink error packet data.

[0128] In some embodiments, the prediction module 403 is further used to obtain the alarm time period of each alarm based on the alarm data; and map the alarm time period to a preset alarm time point to obtain standard alarm data.

[0129] In some embodiments, the prediction module 403 is also used to obtain the target business indicators related to the underlying characteristics of the link in the target network, and also includes: performing preliminary screening on the business data corresponding to each business indicator, retaining only one group of business data with the same change trend within a preset time period as candidate business data; performing business interference cleaning and noise cleaning on the candidate business data to obtain valid business data.

[0130] In some embodiments, service interference cleaning includes: determining from the service data the links where no packet errors have occurred, and eliminating the service data corresponding to the links where no packet errors have occurred; and / or, determining from the service data the service data where there is interference between uplink and downlink and eliminating it; and / or, determining from the service data the service data that is interfered with by the environment and eliminating it.

[0131] In some embodiments, the link risk prediction device further includes a determination module, which is used to determine the link where the candidate service data is located, and to obtain the correspondence between the link where the candidate service data is located and the cell.

[0132] In some embodiments, the link risk prediction device further includes a risk level determination module, which is used to determine the risk level of the link based on the link risk prediction result.

[0133] The link risk prediction device provided in the embodiment of the present application comprises an acquisition module for acquiring target business indicators related to the underlying characteristics of the link in the target network, a determination module for determining the link degradation threshold based on the target business indicators and a preset business indicator degradation threshold, so that the upper-layer target business indicators are dynamically associated with the link degradation threshold of the underlying link, and a prediction module for predicting the risk of the link in the target network based on the link degradation threshold and the link data of the target network within a preset time period, and obtaining a link risk prediction result, so that the basis for link risk prediction is more comprehensive, not only related to the link data at the link level, but also related to the business indicators at the business level, and can not only accurately predict the future risks of the link, such as issuing alarm information in advance, so that the operation and maintenance personnel can eliminate the risks of each link in the target network in advance, avoid the losses caused by blind operation of the operation and maintenance personnel, but also improve user satisfaction.

[0134] Thirdly, refer to Figure 5 , an embodiment of the present application provides an electronic device, comprising:

[0135] One or more processors 501;

[0136] A memory 502, on which one or more programs are stored, and when the one or more programs are executed by one or more processors, the one or more processors implement any one of the above-mentioned link risk prediction methods;

[0137] One or more I / O interfaces 503 are connected between the processor and the memory and are configured to implement information exchange between the processor and the memory.

[0138] Among them, the processor 501 is a device with data processing capabilities, including but not limited to a central processing unit (CPU); the memory 502 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory (FLASH); the I / O interface (read-write interface) 503 is connected between the processor 501 and the memory 502, and can realize information interaction between the processor 501 and the memory 502, including but not limited to a data bus (Bus), etc.

[0139] In some embodiments, the processor 501 , the memory 502 , and the I / O interface 503 are connected to each other through a bus, and further connected to other components of the computing device.

[0140] In a fourth aspect, an embodiment of the present application provides a computer-readable medium having a computer program stored thereon, which implements any of the above-mentioned methods for link risk prediction when the program is executed by a processor.

[0141] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods applied for above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0142] Example embodiments have been claimed herein, and although specific terms are employed, they are used and should be interpreted only in a general illustrative sense and not for limiting purposes. In some instances, it will be apparent to those skilled in the art that, unless otherwise expressly noted, features, characteristics, and / or elements described in conjunction with a particular embodiment may be used alone or in combination with features, characteristics, and / or elements described in conjunction with other embodiments. Therefore, those skilled in the art will appreciate that various changes in form and detail may be made without departing from the scope of the present application as set forth in the appended claims.

Claims

1. A link risk prediction method, characterized in that: include: Obtain target service indicators related to the underlying characteristics of the link in the target network; Determining a link degradation threshold based on the target service indicator and a preset service indicator degradation threshold; The risk of the link in the target network is predicted based on the link degradation threshold and the link data of the target network within a preset time period to obtain a link risk prediction result.

2. The method according to claim 1, characterized in that The obtaining of target service indicators related to the underlying link characteristics in the target network includes: Obtaining multiple fitting points based on the mean value of each candidate business indicator within multiple valid intervals and the interval value of the valid intervals; Using a fitting method to fit the fitting point corresponding to each of the candidate business indicators, to obtain a fitting result corresponding to each of the candidate business indicators; Determining an association parameter between the candidate service indicator and the packet error rate based on the fitting result; The target service indicator is determined based on the association parameter between the candidate service indicator and the packet error rate and a preset parameter threshold.

3. The method according to claim 2, characterized in that The association parameters include significance level and Pearson correlation coefficient, and the parameter thresholds include significance threshold and Pearson threshold.

4. The method according to claim 3, characterized in that The determining the target service indicator based on the association parameter between the candidate service indicator and the packet error rate and a preset parameter threshold comprises: When the significance level is less than the significance threshold and the Pearson correlation coefficient is greater than the Pearson threshold, the candidate business indicator is determined as the target business indicator.

5. The method according to claim 2, characterized in that: The determining the link degradation threshold based on the target service indicator and the preset service indicator degradation threshold includes: Obtaining a link packet error threshold corresponding to each of the target service indicators based on the fitting result; The minimum value of the link packet error thresholds corresponding to the service indicator degradation threshold is used as the link degradation threshold.

6. The method according to claim 2, characterized in that The effective interval is a packet error interval when the number of packet errors within the packet error interval is greater than a preset packet error threshold.

7. The method according to any one of claims 1 to 6, characterized in that: The link data includes link error packet data and alarm data; The predicting the risk of the link in the target network based on the link degradation threshold and the link data of the target network within a preset time period to obtain a link risk prediction result includes: The link degradation threshold and the link data of the target network within a preset time period are input into a risk prediction model, the risk of the link in the target network is predicted, and a link risk prediction result is obtained.

8. The method according to claim 7, characterized in that The risk prediction model is trained by the following steps: Performing standardization processing on the link error packet data to obtain a standard link error packet sample; Standardize the alarm data to obtain standard alarm samples; The risk prediction model is inputted for training based on the standard link error packet samples and the standard alarm samples, and after the risk prediction model converges, the trained risk prediction model is obtained.

9. The method according to claim 8, characterized in that The link degradation threshold includes an uplink degradation threshold and a downlink degradation threshold, and the link error packet data includes uplink error packet data and downlink error packet data; The step of performing standardization processing on the link error packet data sample to obtain standard link error packet data includes: The uplink packet error data in the link packet error data is processed based on the uplink degradation threshold using a packet error rate standardization processing function to obtain standard uplink packet error data, and the downlink packet error data in the link packet error data is processed based on the downlink degradation threshold to obtain standard downlink packet error data.

10. The method according to claim 9, characterized in that The packet error rate normalization processing function is: Among them, i represents the time series, x i represents the uplink error packet data or downlink error packet data at the i-th moment, y i represents the standard uplink error packet data or the standard downlink error packet data, and T represents the link degradation threshold.

11. The method according to claim 8, characterized in that The step of performing standard processing on the alarm data to obtain standard alarm data includes: Obtaining an alarm time period for each alarm based on the alarm data; The alarm time period is mapped to a preset alarm time point to obtain the standard alarm data.

12. The method according to claim 1, characterized in that Before obtaining the target service indicator related to the link bottom layer characteristics in the target network, the method further includes: Perform a preliminary screening of the business data corresponding to each business indicator, and retain only one group of business data with the same change trend within a preset time period as candidate business data; The candidate business data is cleaned to obtain valid business data.

13. The method according to claim 12, characterized in that The cleaning includes one or more of service interference cleaning and noise cleaning.

14. The method according to claim 13, characterized in that The business interference cleaning includes one or more of the following methods: Determine a link where no packet error occurs from the service data, and remove the service data corresponding to the link where no packet error occurs; Determine the service data having interference between uplink and downlink from the service data and remove the service data; The business data affected by the environment is determined from the business data and removed.

15. The method according to claim 12, characterized in that After the business data corresponding to each business indicator is initially screened and only one group of business data with the same change trend within a preset time period is retained as candidate business data, the method further includes: Determine the link where the candidate service data is located; Obtain the correspondence between the link where the candidate service data is located and the cell.

16. The method according to claim 1, characterized in that After predicting the risk of the link in the target network based on the link degradation threshold and the link data of the target network within a preset time period and obtaining the link risk prediction result, the method further includes: The risk level of the link is determined based on the link risk prediction result.

17. A link risk prediction device, characterized in that: include: An acquisition module, used to acquire target service indicators related to the underlying characteristics of the link in the target network; A determination module, configured to determine a link degradation threshold based on the target service indicator and a preset service indicator degradation threshold; The prediction module is used to predict the risk of the link in the target network based on the link degradation threshold and the link data of the target network within a preset time period to obtain a link risk prediction result.

18. An electronic device, characterized in that: include: one or more processors; A storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the link risk prediction method according to any one of claims 1 to 16; One or more I / O interfaces are connected between the processor and the memory and are configured to implement information interaction between the processor and the memory.

19. A computer-readable medium having a computer program stored thereon, wherein when the program is executed by a processor, the link risk prediction method according to any one of claims 1 to 16 is implemented.

Images