Method, system and device for detecting poor quality

Abstract

The invention provides a poor quality detection method, system and device, and the method comprises the steps: obtaining the time delay data of a to-be-detected device, carrying out the data cleaning processing of the time delay data, and obtaining the target time delay data; performing data clustering processing on the target time delay data according to a pre-trained clustering model to obtain a preset number of target time delay center values; performing subtraction processing on each target time delay center value and a reference time delay center value to obtain a calculation result; and under the condition that all the calculation results are greater than a first preset threshold value, determining that the to-be-detected equipment is poor-quality equipment. Compared with the prior art, the embodiment of the invention performs poor quality detection on the time delay data by collecting the time delay data of the to-be-detected equipment, through the pre-trained clustering model and the pre-acquired reference time delay center value, can predict whether the 5G-CPE terminal has poor quality in advance, reduces service interruption caused by poor quality of the 5G-CPE terminal, and improves the service quality of the 5G-CPE terminal. The customer complaints are further reduced, and the use experience of the customer on the 5G private network is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to a method, system and device for quality difference detection. Background Art

[0002] With the popularization of 5G private networks, more and more users have begun to choose 5G-Customer Premise Equipment (5G-CPE) for data transmission. However, in previous 5G private network failures, 5G-CPE terminal abnormalities accounted for more than 90%. Poor quality failures of 5G private network terminals have become a key factor affecting the quality of 5G private networks and have seriously affected users' usage perception.

[0003] In related technologies, when 5G private network users have poor user experience, especially when customers with higher SLA levels have already used it in 7*24 hours of production, customers will complain to relevant departments at any time, and relevant maintenance personnel will manually troubleshoot 5G-CPE terminal failures based on their own accumulated experience. They will further determine whether it is a poor quality terminal and perform corresponding replacement, restart, etc.

[0004] In the existing 5G-CPE terminal poor quality identification technology solution, customers file relevant fault complaints, and maintenance personnel rely on experience to make fault judgments, which requires a large amount of professional personnel, material and financial resources for 7*24 hours monitoring. On the other hand, due to the complex environment of 5G private networks, the usage scenarios, equipment types, and performance of 5G-CPE terminals of various enterprises are different, resulting in great differences in the collected 5G-CPE terminal equipment operation data. Multiple types of data are mixed together, which makes it difficult to manually judge poor quality terminals. It is impossible to quickly locate poor quality in a short time, and the timeliness is low. At the same time, due to the inability to detect poor quality problems in a timely manner, customers' expectations for the stability of 5G private networks are reduced, resulting in a poor user experience. Summary of the invention

[0005] The present disclosure provides a method, system and device for quality difference detection. Its main purpose is to solve the problem in the prior art that manual quality judgment cannot quickly locate the quality difference in a short time, and the timeliness is low. At the same time, due to the inability to detect the quality difference problem in time, the customer's expectation of the stability of the 5G private network is reduced, resulting in a poor user experience.

[0006] According to a first aspect of the present disclosure, a method for quality difference detection is provided, comprising:

[0007] Obtaining delay data of the device to be detected, and performing data cleaning processing on the delay data to obtain target delay data;

[0008] Performing data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model includes the target parameter value;

[0009] Subtract each target delay center value from the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data through the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model;

[0010] In the case where all the calculation results are greater than the first preset threshold, it is determined that the device to be detected is a poor quality device.

[0011] Optionally, before acquiring the delay data of the device to be detected, the method further includes:

[0012] Acquire historical delay data of the device to be detected, and perform data cleaning on the historical delay data to obtain the sample delay data; wherein the historical delay data is the delay data of the device to be detected within a preset time period stored in advance;

[0013] Based on the clustering model to be trained, the sample delay data is distributed and processed according to the data type to obtain the cluster center point corresponding to each type of the sample delay data; wherein the data type is the data type of the sample delay data;

[0014] Based on the clustering model to be trained, repeatedly calculating the cluster center point corresponding to each type of the sample delay data until the change of the cluster center point is less than a preset change threshold, thereby obtaining the pre-trained clustering model, wherein the pre-trained clustering model includes the target parameter value;

[0015] The sample delay data is clustered using the pre-trained clustering model to obtain the reference delay center value.

[0016] Optionally, performing data clustering processing on the target delay data according to a pre-trained clustering model to obtain a target delay center value includes:

[0017] Performing data conversion processing on the target delay data to obtain a target delay vector;

[0018] Performing data adjustment processing on the target delay vector according to the target parameter value to obtain a target data cluster; wherein the number of the target data clusters is the same as the target parameter value;

[0019] Data clustering is performed based on the target data cluster to obtain the target delay center value.

[0020] Optionally, after determining that the device to be detected is a poor-quality device, the method further includes:

[0021] If among all the calculation results, there are calculation results greater than or equal to the preset number threshold and greater than or equal to the second preset threshold, it is determined that the quality difference type of the device to be detected is a special quality difference; wherein the number of the calculation results is the same as the target parameter value;

[0022] If among all the calculation results, there are calculation results greater than or equal to the preset number threshold and less than the second preset threshold, it is determined that the quality difference type of the device to be detected is general quality difference.

[0023] Optionally, after subtracting the target delay center value from the reference delay center value to obtain a calculation result, the method further includes:

[0024] When all the calculation results are less than or equal to the first preset threshold, it is determined that the device to be detected is a normal device.

[0025] If among all the calculation results, there are calculation results greater than or equal to the preset number threshold and greater than or equal to the third preset threshold, it is determined that the quality difference type of the device to be detected is normal operation;

[0026] If all the calculation results are less than a fourth preset threshold, determining that the quality difference type of the device to be detected is a service interruption;

[0027] If among all the calculation results, there is only one calculation result that is smaller than the fifth preset threshold, it is determined that the quality difference type of the device to be detected is that an interruption has occurred.

[0028] Optionally, the method further includes:

[0029] Acquire device information of the device to be detected, and generate information to be sent according to the device information and the quality difference type of the device to be detected;

[0030] The information to be sent is sent to a preset terminal device.

[0031] According to a second aspect of the present disclosure, a system for quality difference detection is provided, comprising: a probe server, a probe device,

[0032] The probe device is used to obtain the delay data of the device to be detected, and perform data cleaning processing on the delay data to obtain the target delay data;

[0033] The probe server is used to perform data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model contains the target parameter value;

[0034] The probe server is also used to subtract each target delay center value from the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data through the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model;

[0035] The probe server is further configured to determine that the device to be detected is a poor-quality device when all the calculation results are greater than a first preset threshold.

[0036] Optionally, the system further includes: a 5G-CPE terminal, a SIM card, a base station device, a core network switch, and a core device.

[0037] The 5G-CPE terminal has the SIM card built in, and is connected and interoperable with the base station device through the 5G air interface; the base station device is connected to the core network switch through the transmission network;

[0038] The SIM card is assigned a fixed IP through the core device, the core network switch communicates data with the 5G-CPE terminal through the fixed IP, and the core network switch is also interconnected with the probe device through an Ethernet optical port optical fiber;

[0039] The probe device is connected to the probe server via a network management network;

[0040] The probe device is also used to determine the delay data by exchanging information with the 5G-CPE terminal;

[0041] The probe server is also used to store the delay data.

[0042] Optionally, the system further comprises: an Internet SMS platform,

[0043] The Internet SMS platform is used to obtain device information of the device to be detected, and generate information to be sent according to the device information and the quality difference type of the device to be detected;

[0044] The Internet SMS platform is also used to send the information to be sent to a preset terminal device.

[0045] According to a third aspect of the present disclosure, a device for quality difference detection is provided, comprising:

[0046] A first acquisition unit is used to acquire the delay data of the device to be detected, and perform data cleaning processing on the delay data to obtain target delay data;

[0047] A clustering unit, configured to perform data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model includes the target parameter value;

[0048] A first calculation unit is used to perform a subtraction process on each of the target delay center values ​​and the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data by the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model;

[0049] The determination unit is used to determine that the device to be detected is a poor quality device when all the calculation results are greater than a first preset threshold.

[0050] Optionally, the first acquisition unit is further used to acquire historical delay data of the device to be detected, and perform data cleaning processing on the historical delay data to obtain the sample delay data; wherein the historical delay data is the delay data of the device to be detected within a preset time period stored in advance;

[0051] The device further includes: an allocating unit, configured to allocate and process the sample delay data according to data types based on the clustering model to be trained, so as to obtain a cluster center point corresponding to each type of the sample delay data; wherein the data type is the data type of the sample delay data;

[0052] A second calculation unit is used to repeatedly calculate the class center point corresponding to each type of the sample delay data based on the clustering model to be trained, until the change of the class center point is less than a preset change threshold, so as to obtain the pre-trained clustering model, wherein the pre-trained clustering model contains the target parameter value;

[0053] The clustering unit is further used to perform clustering processing on the sample delay data through the pre-trained clustering model to obtain the reference delay center value.

[0054] Optionally, the clustering unit includes:

[0055] A conversion module, used for performing data conversion processing on the target delay data to obtain a target delay vector;

[0056] an adjustment module, configured to perform data adjustment processing on the target delay vector according to the target parameter value to obtain a target data cluster; wherein the number of the target data clusters is the same as the target parameter value;

[0057] A clustering module is used to perform data clustering processing based on the target data cluster to obtain the target delay center value.

[0058] Optionally, the determining unit is further configured to:

[0059] When among all the calculation results, there are calculation results greater than or equal to the preset number threshold and greater than or equal to the second preset threshold, determining that the quality difference type of the device to be detected is a special quality difference; wherein the number of the calculation results is the same as the target parameter value;

[0060] When among all the calculation results, there are calculation results greater than or equal to the preset number threshold and less than the second preset threshold, it is determined that the quality difference type of the device to be detected is general quality difference.

[0061] Optionally, the determining unit is further configured to:

[0062] In the case where all the calculation results are less than or equal to the first preset threshold, determining that the device to be detected is a normal device;

[0063] When among all the calculation results, there are calculation results greater than or equal to the preset number threshold and greater than or equal to the third preset threshold, determining that the quality difference type of the device to be detected is normal operation;

[0064] When all the calculation results are less than a fourth preset threshold, determining that the quality difference type of the device to be detected is a service interruption;

[0065] When, among all the calculation results, there is only one calculation result that is smaller than a fifth preset threshold, it is determined that the quality difference type of the device to be detected is that an interruption has occurred.

[0066] Optionally, the device further comprises:

[0067] A second acquisition unit, configured to acquire device information of the device to be detected, and generate information to be sent according to the device information and the quality difference type of the device to be detected;

[0068] The sending unit is used to send the information to be sent to a preset terminal device.

[0069] According to a fourth aspect of the present disclosure, there is provided an electronic device, including:

[0070] at least one processor; and

[0071] a memory communicatively connected to the at least one processor; wherein,

[0072] 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 described in the first aspect.

[0073] According to a fifth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method described in the first aspect.

[0074] According to a sixth aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0075] The method, system and device for quality difference detection provided by the present disclosure obtain delay data of a device to be detected, and perform data cleaning processing on the delay data to obtain target delay data; perform data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model includes the target parameter value; each of the target delay center values ​​is subtracted from the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering sample delay data using the pre-trained clustering model, and the sample delay data is training data for the pre-trained clustering model; when all the calculation results are greater than a first preset threshold, the device to be detected is determined to be a quality difference device. Compared with the related technology, the embodiment of the present disclosure collects the delay data of the device to be detected, and performs quality difference detection on the collected delay data through a pre-trained clustering model and a pre-acquired reference delay center value. It can predict in advance whether the 5G-CPE terminal has poor quality, reduce business interruptions caused by poor quality of the 5G-CPE terminal, further reduce customer complaints, and improve customer experience of using the 5G private network.

[0076] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0078] Figure 1 A flowchart of a method for quality difference detection provided by an embodiment of the present disclosure;

[0079] Figure 2 A schematic diagram of a flow chart of a clustering model training method provided in an embodiment of the present disclosure;

[0080] Figure 3 A schematic diagram of a K value training process provided by an embodiment of the present disclosure;

[0081] Figure 4 A flowchart of another quality difference detection method provided by an embodiment of the present disclosure;

[0082] Figure 5 A schematic diagram of a device for quality difference detection provided by an embodiment of the present disclosure;

[0083] Figure 6 A schematic diagram of the structure of a quality difference detection system provided by an embodiment of the present disclosure;

[0084] Figure 7 A schematic diagram of the structure of a device for quality difference detection provided by an embodiment of the present disclosure;

[0085] Figure 8 A schematic diagram of the structure of another device for quality difference detection provided by an embodiment of the present disclosure;

[0086] Fig. 9 A schematic block diagram of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0087] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0088] The overall architecture of the disclosed embodiment is divided into three parts, including at least: a 5G-CPE terminal, a detection and analysis system, a 5G private network system and a text message forwarding system.

[0089] Among them, the 5G-CPE terminal at least includes: various types of CPE terminal equipment, SIM card configuration fixed IP, and soft probe program; the detection and analysis system at least includes: probe server, probe device, computer storage medium, K-means algorithm program; the 5G private network system at least includes: base station equipment, transmission network, core network switch, 5GC (core equipment); the SMS forwarding system at least includes: Internet SMS platform, mobile phone to receive warning SMS.

[0090] The method, system and device for quality difference detection according to the embodiments of the present disclosure are described below with reference to the accompanying drawings.

[0091] Figure 1 A flowchart of a method for quality difference detection provided in an embodiment of the present disclosure is provided.

[0092] like Figure 1 As shown, the method comprises the following steps:

[0093] Step 101, acquiring delay data of a device to be detected, and performing data cleaning processing on the delay data to obtain target delay data.

[0094] In the embodiments of the present disclosure, the delay data needs to be obtained through the message sending time difference between a preset data acquisition device and the device to be detected. The data acquisition device includes but is not limited to: a probe device, etc., and the device to be detected includes but is not limited to: a 5G-CPE terminal, etc. Specifically, the embodiments of the present disclosure do not impose any restrictions on the device to be detected and the data acquisition device.

[0095] Regarding the acquisition of the delay data, if the data collection device is a probe device and the device to be detected is a 5G-CPE terminal as an example, the collection of delay data includes: the probe device has a built-in dial-up collection application, and the dial-up collection application sends an ICMP Echo message to the fixed IP of the destination node of the 5G-CPE terminal. After the fixed IP of the destination node receives the ICMP Echo message, it will send an Echo Reply message to the probe device. The probe device calculates the delay data based on whether the EchoReply message is received and the time difference between receiving the Echo Reply and sending the ICMP Echo.

[0096] At the same time, in order to facilitate understanding of the implementation process of the embodiments of the present disclosure, the data acquisition device is a probe device and the device to be detected is a 5G-CPE terminal.

[0097] Among them, the step of data cleaning of the delay data includes: performing data judgment on the delay data through a pre-set outlier judgment and analysis module to obtain abnormal data; wherein the abnormal data is data in the delay data that is significantly different from other delay data, and the other delay data is data in the delay data other than the abnormal data; deleting the abnormal data in the delay data to obtain the target delay data.

[0098] When judging abnormal data, you can flexibly adjust the abnormal judgment value according to the customized abnormal range to determine the abnormal data. By reasonably setting the abnormal judgment value, you can not only improve the customer's usage perception, but also reduce unnecessary maintenance personnel input.

[0099] Step 102, performing data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model includes the target parameter value.

[0100] In the disclosed embodiment, the pre-trained clustering model is to collect sample delay data of 5G-CPE terminals (devices to be tested) in different scenarios, and perform K value training on the sample delay data until the initial test centroid no longer changes, so as to determine the initial number of centroids, and then obtain the clustering model (or the clustering algorithm obtained by replacing the original K value in the K-means clustering algorithm with the trained K value), and the trained K value is the target parameter value. Since the collected 5G-CPE terminals not only have different usage scenarios, but also have a particularly large amount of sample delay data, the pre-trained clustering model can find the appropriate number of comparisons between the reference delay center value and the target delay center value, thereby achieving the purpose of optimizing the number of clusters and improving computing efficiency.

[0101] Among them, when calculating the target delay center value, the target delay data will be adjusted to a cluster of target parameter values, each cluster will obtain a target delay center value, and the same number of target delay center values ​​as the target parameter value will be obtained. For example: if the target parameter value is K value, then when calculating the target delay center value, the target delay data will be adjusted to a K value cluster, each cluster will obtain a target delay center value, and there will be K target delay center values ​​in total.

[0102] Step 103, subtract each of the target delay center values ​​from the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data by the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model.

[0103] In the embodiment of the present disclosure, the reference delay center value is obtained by training different K values ​​according to different types, different performances, and different usage scenarios of CPE through the collected sample delay data of 5G-CPE devices (devices to be tested), and using the K-means clustering algorithm to obtain the reference delay center value, that is, the reference delay center value obtained through the pre-trained clustering model.

[0104] The number of the reference delay center values ​​is the same as the number of the target delay center values, that is, when obtaining the reference delay center values, the target delay data will also be adjusted to a cluster of target parameter values, and each cluster will obtain a reference delay center value, and the number of reference delay center values ​​that is the same as the target parameter value will be obtained. That is, the acquisition of the reference delay center value includes but is not limited to: performing data conversion processing on the sample delay data to obtain a reference delay vector; performing data adjustment processing on the reference delay vector according to the target parameter value to obtain a reference data cluster; wherein the number of the reference data clusters is the same as the target parameter value; performing data clustering processing based on the reference data cluster to obtain the reference delay center value.

[0105] When performing the subtraction process of the target delay center value and the reference delay center value, the target delay center value and the reference delay center value of the same number of clusters are compared, for example, the target delay center value corresponding to the first cluster in the target data cluster is subtracted from the reference delay center value corresponding to the first cluster in the reference data cluster. That is, the number of clusters of each target delay center value in the target data cluster and the number of clusters of each reference delay center value in the reference data cluster are obtained respectively, and the target delay center value and the reference delay center value of the same number of clusters are subtracted to obtain the calculation result.

[0106] Step 104: When all the calculation results are greater than a first preset threshold, determine that the device to be detected is a poor quality device.

[0107] In the embodiment of the present disclosure, the first preset threshold is a custom-set value, such as 0, etc. Specifically, the embodiment of the present disclosure does not limit the setting of the first preset threshold.

[0108] According to the data deviation range of each cluster (the calculation results corresponding to each cluster), the quality difference judgment and comparison are performed. When "target delay center value - reference delay center value>0", it is judged as "positive deviation" and the value is output; when "target delay center value - reference delay center value = 0", it is judged as "equal value" and the value is output; when "target delay center value - reference delay center value <0", it is judged as "negative deviation" and the value is output. In order to sort out the number of abnormal delay clusters (that is, whether the calculation results corresponding to each cluster are abnormal), the comparison process can clearly judge the abnormal delay distance, providing a basis for further judgment of quality difference.

[0109] According to the calculation results, further judgment is made that the calculation results simultaneously meet "≥2" conditions, including "target delay center value - reference delay center value = 0", which is judged as "equal value" and "target delay center value - reference delay center value < 0", which is judged as "negative deviation", that is, a negative value. The judgment result is "normal", that is, the device to be tested is a normal device. Otherwise, "target delay center value - reference delay center value > 0", the result is "positive deviation", that is, a positive value, which belongs to "poor quality CPE", that is, the device to be tested is a poor quality device.

[0110] The method for quality difference detection provided by the present disclosure obtains the delay data of the device to be detected, and performs data cleaning processing on the delay data to obtain target delay data; performs data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model contains the target parameter value; each of the target delay center values ​​is subtracted from the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data through the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model; when all the calculation results are greater than the first preset threshold, it is determined that the device to be detected is a quality difference device. Compared with the related technology, the embodiment of the present disclosure collects the delay data of the device to be detected, and performs quality difference detection on the collected delay data through a pre-trained clustering model and a pre-acquired reference delay center value. It can predict in advance whether the 5G-CPE terminal has poor quality, reduce business interruptions caused by poor quality of the 5G-CPE terminal, further reduce customer complaints, and improve customer experience of using the 5G private network.

[0111] In one possible implementation of the embodiment of the present disclosure, in order to find a suitable number of reference center points and target center points for comparison, so as to optimize the number of clusters and improve the operation efficiency, it is necessary to train a suitable clustering model (or clustering algorithm) to complete the clustering of the target delay data, optimize the number of clusters and improve the operation efficiency when performing quality difference detection. Therefore, before obtaining the delay data of the device to be detected, the embodiment of the present disclosure also provides a clustering model training method, such as Figure 2 As shown, Figure 2 A flow chart of a clustering model training method provided in an embodiment of the present disclosure includes:

[0112] Step 201, obtaining the historical delay data of the device to be detected, and performing data cleaning processing on the historical delay data to obtain the sample delay data; wherein the historical delay data is the delay data of the device to be detected within a preset time period that is pre-stored.

[0113] In an embodiment of the present disclosure, the historical delay data is pre-extracted time series data of different types, different performances, and different usage scenarios of the device to be detected, and the historical delay data is stored in the storage medium of the probe server, wherein the extraction time period of the historical delay data, for example: the delay data of the device to be detected within three days, the delay data of the device to be detected within one week, etc. Specifically, the embodiment of the present disclosure does not limit the time period of the historical delay data.

[0114] Regarding the cleaning of the historical delay data, reference may be made to the cleaning of the delay data in the above step 101, so it will not be described in detail here.

[0115] Step 202: Based on the clustering model to be trained, the sample delay data is allocated and processed according to the data type to obtain the cluster center point corresponding to each type of the sample delay data; wherein the data type is the data type of the sample delay data.

[0116] In the disclosed embodiment, the clustering model to be trained is a model structure of user-defined selection, such as K-means clustering algorithm, etc. When training the clustering model to be trained, it is necessary to train the K value according to the sample delay data until the initial centroid no longer changes, so as to determine the initial number of centroids. After that, the original K value in the clustering model to be trained (K-means clustering algorithm) is replaced by the trained K value to obtain a pre-trained clustering model.

[0117] Among them, by iteratively adjusting the centroid to minimize the intra-class distance and maximize the inter-class distance, the sample delay data point is assigned to the nearest class center point to obtain the class center point of the first iteration. When allocating the sample delay data, it is allocated according to the data type of the sample delay data. The data type includes but is not limited to: the type, performance, usage scenario, etc. of the device to be tested corresponding to the sample delay data.

[0118] The class center point refers to the center point of each cluster (each reference data cluster), which is the mean position of all data points in the cluster. In each iteration of the algorithm, the class center point is recalculated to minimize the sum of the distances from the data points in the cluster to its center point.

[0119] The original K value is a parameter specified by the user, that is, a parameter determined by the user according to the type of the sample delay data, indicating how many clusters the data set is to be divided into. Then, the K value is trained based on the original K value to obtain the trained K value. Specifically, regarding the training of the K value, the embodiment of the present disclosure provides a flow chart of the K value training, as shown in FIG. Figure 3 As shown, K value training is performed through the optimal K value training module.

[0120] The initial centroid is the initial center point of each cluster when the algorithm starts to iterate. The selection of the initial centroid will affect the convergence speed of the algorithm and the quality of the final clustering result. Common methods for selecting the initial centroid include random selection, Forgy algorithm, K-means++ algorithm, and methods based on data density distribution. When the sample delay data is distributed and processed according to the data type, the sample delay data is distributed to each corresponding initial centroid according to the data type, and the class center point of the first iteration is calculated.

[0121] Step 203, based on the clustering model to be trained, repeatedly calculate the class center point corresponding to each type of sample delay data until the change of the class center point is less than a preset change threshold, and obtain the pre-trained clustering model, wherein the pre-trained clustering model contains the target parameter value.

[0122] In the disclosed embodiment, the center point of each class is calculated repeatedly, and this process is repeated until the class no longer changes or the maximum number of iterations is reached, so that the pre-trained clustering model can be obtained.

[0123] Step 204: cluster the sample delay data using the pre-trained clustering model to obtain the reference delay center value.

[0124] Specifically, regarding the clustering process of the reference delay center value, reference may be made to the description in the above step 103, so it will not be described in detail here.

[0125] In one implementable manner of the embodiment of the present disclosure, when performing data clustering processing on the target delay data to obtain the target delay center value, it can also be implemented in but not limited to the following manner: performing data conversion processing on the target delay data to obtain a target delay vector; performing data adjustment processing on the target delay vector according to the target parameter value to obtain a target data cluster; wherein the number of the target data clusters is the same as the target parameter value; performing data clustering processing based on the target data cluster to obtain the target delay center value.

[0126] In an embodiment of the present disclosure, the delay data (target delay data) stored in the probe server is extracted and reshaped into a vector form (to obtain a target delay vector), and the data is adjusted into a K-value cluster trained in different scenarios, i.e., a target data cluster. After that, the target delay center value corresponding to each cluster can be directly clustered.

[0127] In one implementable manner of the embodiment of the present disclosure, after determining that the device to be detected is a poor quality device, it is necessary to determine the quality difference type of the device to be detected, so as to notify relevant maintenance personnel according to the quality difference type and quickly locate the quality difference. Therefore, in order to quickly locate the quality difference and improve efficiency, the following method may also be used but is not limited to: if among all the calculation results, there are calculation results greater than or equal to a preset number threshold and greater than or equal to a second preset threshold, then the quality difference type of the device to be detected is determined to be special quality difference; wherein the number of the calculation results is the same as the target parameter value; if among all the calculation results, there are calculation results greater than or equal to the preset number threshold and less than the second preset threshold, then the quality difference type of the device to be detected is determined to be general quality difference.

[0128] In the embodiment of the present disclosure, the preset quantity threshold is a threshold of a custom setting, for example: half of the target parameter value, etc., the second preset threshold is a threshold of a custom setting, for example: 5, etc., the second preset threshold can be adjusted according to the actual scene on site. If the latency requirement is high, the second preset threshold can be set lower. Specifically, the embodiment of the present disclosure does not limit the preset quantity threshold and the second preset threshold.

[0129] According to the value in the "poor quality CPE" (poor quality equipment) type defined by the analysis result, the poor quality type is further determined. If "poor quality CPE ≥ 1 / 2" (indicating half of the target parameter value), that is, when the difference between the reference delay center value and the target delay center value is compared, half or more of the numbers are positive and "comparison value (difference) ≥ 5" (in all calculation results, more than or equal to 1 / 2 of the calculation results are comparison values ​​(difference) ≥ 5), then "particularly poor quality" is output. If "poor quality CPE ≥ 1 / 2", that is, when the difference between the reference delay center value and the target delay center value is compared, half or more of the numbers are within the range of "0 < comparison value (difference) < 5" (in all calculation results, more than or equal to 1 / 2 of the calculation results are 0 < comparison value (difference) < 5), then "general poor quality" is output, where 1 / 2 is the preset quantity threshold and 5 is the second preset threshold.

[0130] In one implementable manner of the embodiments of the present disclosure, associated with the above embodiments, the device to be tested is not necessarily a poor quality device, but may also be a normal device. After determining that the device to be tested is a poor quality device, it is necessary to determine the poor quality type of the device to be tested. Similarly, after determining that the device to be tested is a normal device, it is also necessary to determine the normal type of the device to be tested. Specifically, but not limited to, the following method may be used: when all of the calculation results are less than or equal to the first preset threshold, the device to be tested is determined to be a normal device; if among all of the calculation results, there are calculation results greater than or equal to the preset number threshold and greater than or equal to the third preset threshold, then the poor quality type of the device to be tested is determined to be normal operation; if all of the calculation results are less than the fourth preset threshold, then the poor quality type of the device to be tested is determined to be a business interruption; if among all of the calculation results, there is only one calculation result less than the fifth preset threshold, then the poor quality type of the device to be tested is determined to have been interrupted.

[0131] The preset quantity threshold is a threshold of a custom setting, for example: half of the target parameter value, etc. The third preset threshold is a threshold of a custom setting, for example: -9, etc. The fourth preset threshold and the fifth preset threshold can be the same as or different from the third preset threshold, for example: both are -9, or the fourth preset threshold is -12, the fifth preset threshold is -10, etc. Specifically, the settings of the preset quantity threshold, the third preset threshold, the fourth preset threshold and the fifth preset threshold can be determined according to actual conditions, and the embodiments of the present disclosure are not limited.

[0132] If "quality difference CPE ≥ 1 / 2" (indicates half of the target parameter value), that is, when the difference between the reference delay center value and the target delay center value is compared, half or more of the numbers are within the range of "-9 ≤ comparison value (difference) < 0" (of all calculation results, more than or equal to 1 / 2 of the calculation results are -9 ≤ comparison value (difference) < 0), then the output is "normal operation" and the next round of 1-hour granularity quality difference analysis is performed; if it is judged that after the difference between the reference center value and the target center value is compared, there is only one cluster that meets the "comparison value (difference)" range. <-12" condition (among all calculation results, there is only one calculation result with comparison value (difference) <-12), then output "interruption occurred"; if it is judged that after comparing the difference between the reference center value and the target center value, all values ​​"comparison value (difference) <-10" (all calculation results are comparison value (difference) <-10), then output "business interruption", where 1 / 2 is the preset quantity threshold, -9 is the third preset threshold, -12 is the fourth preset threshold, and -10 is the fifth preset threshold.

[0133] In one implementable method of the embodiment of the present disclosure, it is necessary to send the quality difference type of the device to be detected and the device information of the device to be detected to the maintenance personnel, so that the maintenance personnel can discover the problem before the user and realize rapid quality difference positioning and identification. Therefore, in order to provide data support for on-site front-line maintenance personnel, the following method can also be adopted but not limited to: obtain the device information of the device to be detected, and generate the information to be sent according to the device information and the quality difference type of the device to be detected; send the information to be sent to a preset terminal device.

[0134] In the embodiments of the present disclosure, the device information includes but is not limited to: the SIM card IP address assigned by the 5GC (core device) to the poor quality CPE (device to be detected), the collected poor quality CPE name (name of the device to be detected), etc. The poor quality type of the device to be detected includes but is not limited to: "particularly poor quality", "generally poor quality", "normal operation", "interruption occurred" and "service interruption", etc.

[0135] Among them, according to the output results of "poor quality CPE" type "particularly poor quality", "generally poor quality", "normal operation", "interruption occurred" and "service interruption", the probe server further generates poor quality SMS text, including "SIM card IP address assigned to poor quality CPE by 5GC" and "collected poor quality CPE name", and sends it to the Internet SMS platform through the network management network. The warning SMS text sent by the probe server received by the Internet SMS platform is converted into a short message and sent to the target maintenance personnel or target users. It can improve the accuracy of poor quality positioning and reduce the cost of manpower input.

[0136] Further, based on the above embodiments, the present disclosure provides a flowchart of another method for quality difference detection, such as Figure 4 As shown, this process method analysis can reduce the input of analysts, provide data support for on-site maintenance personnel, and realize the ability to quickly judge the poor quality of 5G-CPE terminals.

[0137] In one possible implementation of the embodiment of the present disclosure, the embodiment of the present disclosure provides a schematic diagram of a device for quality difference detection, such as Figure 5 As shown, the 5G-CPE quality difference is the device for quality difference detection. The training module and clustering algorithm module deployed in the probe server can accurately analyze and output the positioning results of the 5G-CPE quality difference, and promptly send SMS reminders to maintenance personnel or users to extract predicted quality difference terminal problems. It can reduce the analysis time and manpower invested by professional and technical personnel and save labor maintenance costs. Regarding the 5G-CPE analysis mentioned in the embodiments of the present disclosure, the K value can be trained separately for each scene or park without repeated training. Or if there are multiple CPE types from different manufacturers in the same park, data training can be performed according to the manufacturer.

[0138] The present invention also proposes a system for quality difference detection, such as Figure 6 As shown, Figure 6 A schematic diagram of a system for quality difference detection provided by an embodiment of the present disclosure includes: a probe server, a probe device,

[0139] The probe device is used to obtain the delay data of the device to be detected, and perform data cleaning processing on the delay data to obtain the target delay data;

[0140] The probe server is used to perform data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model contains the target parameter value;

[0141] The probe server is also used to subtract each target delay center value from the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data through the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model;

[0142] The probe server is further configured to determine that the device to be detected is a poor-quality device when all the calculation results are greater than a first preset threshold.

[0143] In one possible implementation of the embodiment of the present disclosure, the system further includes: a 5G-CPE terminal, a SIM card, a base station device, a core network switch, and a core device.

[0144] The 5G-CPE terminal has the SIM card built in, and is connected and interoperable with the base station device through the 5G air interface; the base station device is connected to the core network switch through the transmission network;

[0145] The SIM card is assigned a fixed IP through the core device, the core network switch communicates data with the 5G-CPE terminal through the fixed IP, and the core network switch is also interconnected with the probe device through an Ethernet optical port optical fiber;

[0146] The probe device is connected to the probe server via a network management network;

[0147] The probe device is also used to determine the delay data by exchanging information with the 5G-CPE terminal;

[0148] The probe server is also used to store the delay data.

[0149] In one possible implementation of the embodiment of the present disclosure, the system further includes: an Internet short message platform,

[0150] The Internet SMS platform is used to obtain device information of the device to be detected, and generate information to be sent according to the device information and the quality difference type of the device to be detected;

[0151] The Internet SMS platform is also used to send the information to be sent to a preset terminal device.

[0152] Among them, see Figure 6 , 5G-CPE terminals include: various types of CPE terminal equipment, SIM card configuration fixed IP, soft probe program; the detection and analysis system includes: probe server, probe device, computer storage medium, K-means algorithm program; 5G private network system includes: base station equipment, transmission network, core network switch, 5GC (core equipment); SMS forwarding system includes: Internet SMS platform, mobile phone receiving warning SMS.

[0153] The quality difference detection method of 5G private network CPE terminal and the connection relationship between devices, algorithm analysis process, and warning message processing method are as follows:

[0154] The connection relationship between devices includes: the 5G-CPE terminal has a built-in SIM card, and the SIM card is assigned a fixed IP through 5GC; the 5G-CPE terminal is connected to the base station equipment through the 5G air interface and sends the delay data to the base station; the base station equipment transmits the delay data to the core network switch by connecting to the transmission network; the core network switch and the probe device are interconnected by Ethernet optical fiber, and the fixed IP route of the SIM card of the CPE device is opened by configuring IP routing on the core network switch to realize data intercommunication; the probe device is connected to the probe server through the network management network to forward and monitor the IP delay data of the SIM card IP of the CPE device; the probe server collects delay data through the probe device connected to the network management network, including: the probe device has a built-in dialing collection application, which sends an ICMP Echo message to the fixed IP of the destination node of the 5G-CPE terminal. After receiving the Echo message, the destination node IP will send an Echo Reply message to the sending probe device. The sending end calculates the detection delay data based on whether the Reply message is received and the time difference between receiving the Echo Reply and sending the Echo. The delay data is stored in the storage medium of the probe server; the optimal K value training module of the 5G-CPE quality difference identification and analysis device deployed by the probe server calls the delay data stored in the probe server at a minute granularity period and performs optimal K value training. Due to different CPE types, different performances, and different usage scenarios, typical reference center value case data is generated through clustering delay characteristics and training, and then combined with the actual characteristics of the customer's target delay, differentiated and accurate matching is used to formulate the optimal center point strategy to improve customer perception. The training K value results are output to the outlier determination and analysis module for processing; after the outlier determination and analysis module excludes the outliers, the K-means clustering reference and target center value depth deviation result analysis is performed; after the quality difference analysis and determination result module is analyzed and determined, the next round of analysis is performed if the result type is normal, otherwise a quality difference SMS is generated to notify the maintenance personnel to restart or replace the process.

[0155] In summary, the embodiments of the present disclosure can achieve the following effects:

[0156] 1. The disclosed embodiment collects the delay data of the device to be tested, and performs quality difference detection on the collected delay data through a pre-trained clustering model and a pre-acquired reference delay center value. This can predict in advance whether the 5G-CPE terminal has quality poor, reduce service interruptions caused by poor quality of the 5G-CPE terminal, further reduce customer complaints, and improve customer experience of using the 5G private network.

[0157] 2. The disclosed embodiment obtains the output result of poor quality through a fast and simple method for collecting the latency phenomenon of 5G-CPE terminals, an unsupervised learning clustering algorithm, a square sum difference algorithm, etc., and notifies the operation and maintenance personnel to replace or restart the 5G-CPE terminals based on the output results. This can predict in advance whether the 5G-CPE terminals are of poor quality, reduce business interruptions caused by poor quality of 5G-CPE terminals, further reduce customer complaints, and improve customers' overall satisfaction with the stability of 5G private networks and 5G-A networks.

[0158] 3. The disclosed embodiment deploys a probe device to send a detection task to the destination IP address of the 5G-CPE terminal, and detects the destination IP address of the 5G-CPE terminal node based on the customer scenario and optionally adjustable sensitivity parameters, thereby achieving the purpose of enhancing customers' perception of 5G private network services, reducing manpower and operation and maintenance cost investment, discovering problems before users, and quickly locating and identifying quality differences.

[0159] Corresponding to the above-mentioned quality difference detection method, the present invention also provides a quality difference detection device. Since the device embodiment of the present invention corresponds to the above-mentioned method embodiment, the details not disclosed in the device embodiment can be referred to the above-mentioned method embodiment, and will not be repeated in the present invention.

[0160] Figure 7 A schematic diagram of a device for quality difference detection provided by an embodiment of the present disclosure is shown in FIG. Figure 7 As shown, including:

[0161] A first acquisition unit 71 is used to acquire the delay data of the device to be detected, and perform data cleaning processing on the delay data to obtain target delay data;

[0162] A clustering unit 72 is used to perform data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model contains the target parameter value;

[0163] The first calculation unit 73 is used to perform subtraction processing on each of the target delay center values ​​and the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data by the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model;

[0164] The determination unit 74 is configured to determine that the device to be detected is a poor-quality device when all the calculation results are greater than a first preset threshold.

[0165] The device for quality difference detection provided by the present disclosure obtains the delay data of the device to be detected, and performs data cleaning processing on the delay data to obtain target delay data; performs data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model contains the target parameter value; each of the target delay center values ​​is subtracted from the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data through the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model; when all the calculation results are greater than the first preset threshold, it is determined that the device to be detected is a quality difference device. Compared with the related technology, the embodiment of the present disclosure collects the delay data of the device to be detected, and performs quality difference detection on the collected delay data through a pre-trained clustering model and a pre-acquired reference delay center value. It can predict in advance whether the 5G-CPE terminal has poor quality, reduce business interruptions caused by poor quality of the 5G-CPE terminal, further reduce customer complaints, and improve customer experience of using the 5G private network.

[0166] Further, in a possible implementation of the embodiment of the present disclosure, the first acquisition unit 71 is further used to acquire the historical delay data of the device to be detected, and perform data cleaning processing on the historical delay data to obtain the sample delay data; wherein the historical delay data is the delay data of the device to be detected within a preset time period stored in advance;

[0167] like Figure 8 As shown, the device further includes: an allocation unit 75, which is used to allocate and process the sample delay data according to the data type based on the clustering model to be trained, and obtain the class center point corresponding to each type of the sample delay data; wherein the data type is the data type of the sample delay data;

[0168] A second calculation unit 76 is used to repeatedly calculate the cluster center point corresponding to each type of the sample delay data based on the cluster model to be trained, until the change of the cluster center point is less than a preset change threshold, to obtain the pre-trained cluster model, wherein the pre-trained cluster model contains the target parameter value;

[0169] The clustering unit 72 is further configured to perform clustering processing on the sample delay data using the pre-trained clustering model to obtain the reference delay center value.

[0170] Furthermore, in a possible implementation of the embodiment of the present disclosure, as Figure 8 As shown, the clustering unit 72 includes:

[0171] The conversion module 721 is used to perform data conversion processing on the target delay data to obtain a target delay vector;

[0172] An adjustment module 722, configured to perform data adjustment processing on the target delay vector according to the target parameter value to obtain a target data cluster; wherein the number of the target data clusters is the same as the target parameter value;

[0173] The clustering module 723 is used to perform data clustering processing based on the target data cluster to obtain the target delay center value.

[0174] Furthermore, in a possible implementation of the embodiment of the present disclosure, the determining unit 74 is further configured to:

[0175] When among all the calculation results, there are calculation results greater than or equal to the preset number threshold and greater than or equal to the second preset threshold, determining that the quality difference type of the device to be detected is a special quality difference; wherein the number of the calculation results is the same as the target parameter value;

[0176] When among all the calculation results, there are calculation results greater than or equal to the preset number threshold and less than the second preset threshold, it is determined that the quality difference type of the device to be detected is general quality difference.

[0177] Furthermore, in a possible implementation of the embodiment of the present disclosure, the determining unit 74 is further configured to:

[0178] In the case where all the calculation results are less than or equal to the first preset threshold, determining that the device to be detected is a normal device;

[0179] When among all the calculation results, there are calculation results greater than or equal to the preset number threshold and greater than or equal to the third preset threshold, determining that the quality difference type of the device to be detected is normal operation;

[0180] When all the calculation results are less than a fourth preset threshold, determining that the quality difference type of the device to be detected is a service interruption;

[0181] When, among all the calculation results, there is only one calculation result that is smaller than a fifth preset threshold, it is determined that the quality difference type of the device to be detected is that an interruption has occurred.

[0182] Furthermore, in a possible implementation of the embodiment of the present disclosure, as Figure 8 As shown, the device also includes:

[0183] A second acquisition unit 77 is used to acquire device information of the device to be detected, and generate information to be sent according to the device information and the quality difference type of the device to be detected;

[0184] The sending unit 78 is used to send the information to be sent to a preset terminal device.

[0185] It should be noted that the above explanation of the method embodiment is also applicable to the device of the embodiment of the present disclosure, and the principle is the same, which is no longer limited in the embodiment of the present disclosure.

[0186] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0187] Fig. 9 A schematic block diagram of an example electronic device 900 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0188] like Fig. 9 As shown, the device 900 includes a computing unit 901, which can perform various appropriate actions and processes according to a computer program stored in a ROM (Read-Only Memory) 902 or a computer program loaded from a storage unit 908 to a RAM (Random Access Memory) 903. In the RAM 903, various programs and data required for the operation of the device 900 can also be stored. The computing unit 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An I / O (Input / Output) interface 905 is also connected to the bus 904.

[0189] A number of components in the device 900 are connected to the I / O interface 905, including: an input unit 906, such as a keyboard, a mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a disk, an optical disk, etc.; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 909 allows the device 900 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0190] The computing unit 901 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, CPU (Central Processing Unit), GPU (Graphic Processing Units), various dedicated AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, DSP (Digital Signal Processor), and any appropriate processor, controller, microcontroller, etc. The computing unit 901 performs the various methods and processes described above, such as the method of quality difference detection. For example, in some embodiments, the method of quality difference detection may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 908. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 900 via the ROM 902 and / or the communication unit 909. When the computer program is loaded into the RAM 903 and executed by the computing unit 901, one or more steps of the method described above may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured to perform the aforementioned quality difference detection method in any other appropriate manner (for example, by means of firmware).

[0191] Various embodiments of the systems and techniques described above herein may be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application Specific Standard Products), SOCs (System On Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that may be executed and / or interpreted on a programmable system including at least one programmable processor that may be a special purpose or general purpose programmable processor that may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0192] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0193] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a RAM, a ROM, an EPROM (Electrically Programmable Read-Only-Memory) or a flash memory, an optical fiber, a CD-ROM (Compact Dis sc Read-Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0194] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball), through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0195] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: LAN (Local Area Network), WAN (Wide Area Network), the Internet, and blockchain networks.

[0196] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server of a distributed system, or a server combined with a blockchain.

[0197] It should be noted that artificial intelligence is a discipline that studies how computers can simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.), and includes both hardware-level and software-level technologies. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, and big data processing; artificial intelligence software technologies mainly include computer vision technology, speech recognition technology, natural language processing technology, as well as machine learning / deep learning, big data processing technology, knowledge graph technology, and other major directions.

[0198] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0199] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for quality difference detection, characterized in that: include: Obtaining delay data of the device to be detected, and performing data cleaning processing on the delay data to obtain target delay data; Performing data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model includes the target parameter value; Subtract each target delay center value from the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data through the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model; In the case where all the calculation results are greater than the first preset threshold, it is determined that the device to be detected is a poor quality device.

2. The method according to claim 1, characterized in that: Before acquiring the delay data of the device to be detected, the method further includes: Acquire historical delay data of the device to be detected, and perform data cleaning on the historical delay data to obtain the sample delay data; wherein the historical delay data is the delay data of the device to be detected within a preset time period stored in advance; Based on the clustering model to be trained, the sample delay data is distributed and processed according to the data type to obtain the cluster center point corresponding to each type of the sample delay data; wherein the data type is the data type of the sample delay data; Based on the clustering model to be trained, repeatedly calculating the cluster center point corresponding to each type of the sample delay data until the change of the cluster center point is less than a preset change threshold, thereby obtaining the pre-trained clustering model, wherein the pre-trained clustering model includes the target parameter value; The sample delay data is clustered using the pre-trained clustering model to obtain the reference delay center value.

3. The method according to claim 1, characterized in that The performing data clustering processing on the target delay data according to the pre-trained clustering model to obtain the target delay center value includes: Performing data conversion processing on the target delay data to obtain a target delay vector; Performing data adjustment processing on the target delay vector according to the target parameter value to obtain a target data cluster; wherein the number of the target data clusters is the same as the target parameter value; Data clustering is performed based on the target data cluster to obtain the target delay center value.

4. The method according to claim 1, characterized in that: After determining that the device to be detected is a poor-quality device, the method further includes: If among all the calculation results, there are calculation results greater than or equal to the preset number threshold and greater than or equal to the second preset threshold, it is determined that the quality difference type of the device to be detected is a special quality difference; wherein the number of the calculation results is the same as the target parameter value; If among all the calculation results, there are calculation results greater than or equal to the preset number threshold and less than the second preset threshold, it is determined that the quality difference type of the device to be detected is general quality difference.

5. The method according to claim 1, characterized in that After subtracting the target delay center value from the reference delay center value to obtain a calculation result, the method further includes: In the case where all the calculation results are less than or equal to the first preset threshold, determining that the device to be detected is a normal device; If among all the calculation results, there are calculation results greater than or equal to the preset number threshold and greater than or equal to the third preset threshold, it is determined that the quality difference type of the device to be detected is normal operation; If all the calculation results are less than a fourth preset threshold, determining that the quality difference type of the device to be detected is a service interruption; If among all the calculation results, there is only one calculation result that is smaller than the fifth preset threshold, it is determined that the quality difference type of the device to be detected is that an interruption has occurred.

6. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: Acquire device information of the device to be detected, and generate information to be sent according to the device information and the quality difference type of the device to be detected; The information to be sent is sent to a preset terminal device.

7. A system for quality difference detection, characterized in that: include: Probe server, probe device, The probe device is used to obtain the delay data of the device to be detected, and perform data cleaning processing on the delay data to obtain the target delay data; The probe server is used to perform data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model contains the target parameter value; The probe server is also used to subtract each target delay center value from the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data through the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model; The probe server is further configured to determine that the device to be detected is a poor-quality device when all the calculation results are greater than a first preset threshold.

8. The system according to claim 7, characterized in that The system also includes: 5G-CPE terminal, SIM card, base station equipment, core network switch, core equipment, The 5G-CPE terminal has the SIM card built in, and is connected and interoperable with the base station device through the 5G air interface; the base station device is connected to the core network switch through the transmission network; The SIM card is assigned a fixed IP through the core device, the core network switch communicates data with the 5G-CPE terminal through the fixed IP, and the core network switch is also interconnected with the probe device through an Ethernet optical port optical fiber; The probe device is connected to the probe server via a network management network; The probe device is also used to determine the delay data by exchanging information with the 5G-CPE terminal; The probe server is also used to store the delay data.

9. The system according to claim 8, characterized in that The system also includes: an Internet short message platform, The Internet SMS platform is used to obtain device information of the device to be detected, and generate information to be sent according to the device information and the quality difference type of the device to be detected; The Internet SMS platform is also used to send the information to be sent to a preset terminal device.

10. A device for quality difference detection, characterized in that: include: A first acquisition unit is used to acquire the delay data of the device to be detected, and perform data cleaning processing on the delay data to obtain target delay data; A clustering unit, configured to perform data clustering processing on the target delay data according to a pre-trained clustering model to obtain a preset number of target delay center values; wherein the number of the target delay center values ​​is the same as the target parameter value, and the pre-trained clustering model includes the target parameter value; A first calculation unit is used to perform a subtraction process on each of the target delay center values ​​and the reference delay center value to obtain a calculation result; wherein the reference delay center value is obtained by clustering the sample delay data by the pre-trained clustering model, and the sample delay data is the training data of the pre-trained clustering model; The determination unit is used to determine that the device to be detected is a poor quality device when all the calculation results are greater than a first preset threshold.

Images