An abnormal switching user determination method, system and device

Abstract

The application relates to an abnormal handover user determination method, system and device. The method comprises the following steps: determining a user to be handed over from users accessing a target cell; obtaining network configuration data of the user to be handed over; inputting the network configuration data into a first deep network model, so that the first deep network model outputs a first probability value of the user to be handed over being a normal handover user and a second probability value of the user to be handed over being an abnormal handover user according to network configuration data features of the normal handover user and network configuration data features of the abnormal handover user respectively; and determining whether the user to be handed over is an abnormal handover user according to the first probability value and the second probability value. In the embodiment of the application, the network configuration data of the user to be handed over is input into the first deep network model, so that the first deep network model automatically determines whether the user to be handed over is an abnormal handover user according to data features of the network configuration data of the user to be handed over.

Description

[Technical Field]

[0001] This invention relates to the field of network cutover, and in particular to a method, system, and device for identifying users during abnormal cutovers. [Background Technology]

[0002] With the continuous expansion of network services, the demand for Passive Optical Network (PON) equipment is also increasing. To improve network service expansion and user experience, PON networks need frequent optimization. Network cutover is often used in PON network optimization. Network cutover refers to migrating some users (the users to be cut over) from PON ports with high network resource utilization to PON ports with lower utilization. During network cutover, there may be discrepancies between the resource data stored on the server and the configuration data corresponding to the actual user connections, leading to cutover failure. This results in unusable services for users whose connections failed, causing user complaints. For example, the server might store user A's PON port as 1, and the corresponding secondary optical splitter as 1. However, in actual wiring, user A is connected to PON port 2, and the corresponding secondary optical splitter is also 2. In this case, performing network cutover on user A based on the resource data stored on the server will result in cutover failure, making the services of users actually connected to secondary optical splitter 1 on PON port 1 unusable. [Summary of the Invention]

[0003] To address the aforementioned problems, embodiments of the present invention provide a method, system, and device for determining abnormal cutover users. Before performing a network cutover, the network configuration data of the user to be cut over is used to determine whether the user is an abnormal cutover user, thereby avoiding cutover failure.

[0004] In a first aspect, embodiments of the present invention provide a method for determining abnormal cutover users, including:

[0005] Identify the users to be cutovered from the users connected to the target cell;

[0006] Obtain the network configuration data of the user to be cut over;

[0007] The network configuration data is input into a first deep network model, so that the first deep network model outputs a first probability value that the user to be cut over is a normal user and a second probability value that is an abnormal user, respectively, based on the network configuration data characteristics of normal cutover users and network configuration data characteristics of abnormal cutover users.

[0008] Based on the first probability value and the second probability value, determine whether the user to be cut over is an abnormal cutover user.

[0009] In this embodiment of the invention, after identifying the user to be cut over, the network configuration data of the user to be cut over is obtained and input into a first deep network model. Then, based on the output of the first deep network model, it is determined whether the user to be cut over is an abnormal cutover user. This reduces the probability of network cutover failure.

[0010] In one possible implementation, the users to be cutover are determined from the users accessing the target cell, including:

[0011] Obtain the load information of each network interface of the target cell, wherein the target cell contains multiple network interfaces, each network interface contains multiple optical splitters, each optical splitter corresponds to multiple users, and the load information contains multiple parameter items;

[0012] Determine the score and weight of each parameter item for each network interface;

[0013] The scores and weights of each parameter item are weighted to obtain the load score for each network interface;

[0014] The network interfaces whose load scores are greater than the first threshold are identified as network interfaces to be cut over.

[0015] The optical splitter containing more than a second threshold number of users is identified from the optical splitters included in the network interface to be cut over as the optical splitter to be cut over.

[0016] The user corresponding to the splitter to be cut is identified as the user to be cut.

[0017] In one possible implementation, the network configuration data is input into a first deep network model, so that the first deep network model outputs a first probability value for the user to be cut over as a normal cutover user and a second probability value for the user to be cut over as an abnormal cutover user, respectively, based on the network configuration data characteristics of normal cutover users and abnormal cutover users, including:

[0018] Based on the network configuration data, construct the network configuration dataset of the users to be cutoverdated;

[0019] Perform fixed difference operation or one-hot encoding operation on the network configuration dataset;

[0020] Based on the calculation results, construct the network configuration data matrix for the users to be cut over;

[0021] The network configuration data matrix is ​​input into the first deep network model so that the first deep network model determines the network configuration data characteristics of the user to be cut over based on the network configuration data matrix, and outputs a first probability value that the user to be cut over is a normal user and a second probability value that is an abnormal user based on the network configuration data characteristics of normal users and abnormal users.

[0022] In one possible implementation, the training process of the first deep network model includes:

[0023] Obtain historical network configuration data, which includes network configuration data labeled with normal cutover users and network configuration data labeled with abnormal cutover users;

[0024] Construct a historical network configuration data matrix based on the historical network configuration data;

[0025] The historical network configuration data matrix is ​​input into the first deep network model for iterative training, so that the first deep network model learns the network configuration data characteristics of normal cutover users and the network configuration data characteristics of abnormal cutover users.

[0026] The iteration ends when the accuracy of the output result of the first deep network model is greater than the third threshold.

[0027] In one possible implementation, the method further includes:

[0028] If the user to be cut over is an abnormal user, then the network configuration data of the user to be cut over is input into the second deep network model, so that the second deep network model outputs the probability value of the user to be cut over as an abnormal problem of each category according to the network configuration data characteristics of different categories of abnormal problems.

[0029] The anomaly category of the user to be cut over is determined based on the probability value of each category of anomaly.

[0030] Based on the category of the abnormal problem, determine the abnormal configuration parameters of the user to be cut over.

[0031] In one possible implementation, the training process of the second deep network model includes:

[0032] Obtain historical abnormal network configuration data, which includes abnormal problem tags labeled with different categories, and each abnormal problem tag corresponds to an abnormal configuration parameter;

[0033] Construct a historical abnormal network configuration data matrix based on the historical abnormal network configuration data;

[0034] The historical anomaly network configuration data matrix is ​​input into the second deep network model for iterative training, so that the second deep network model learns the network configuration data features corresponding to each category of anomalies.

[0035] The iteration ends when the accuracy of the output result of the second deep network model is greater than the fourth threshold.

[0036] In one possible implementation, the method further includes:

[0037] Based on the probability values ​​of the user to be cut over as an abnormal problem of each category output by the second deep network model, determine the abnormal problem category with the smallest probability value;

[0038] Based on the anomaly category with the lowest probability value, determine the configuration parameter with the lowest anomaly probability;

[0039] The repair value of the abnormal configuration parameter is determined based on the value of the configuration parameter with the lowest probability of abnormality.

[0040] The abnormal configuration parameters are repaired based on the repair values.

[0041] Secondly, embodiments of the present invention provide an abnormal cutover user determination system, comprising:

[0042] The determination module is used to identify users to be cut over from the users accessing the target cell;

[0043] The acquisition module acquires the network configuration data of the user to be cutoverdone.

[0044] The deep learning module is used to input the network configuration data into the first deep network model, so that the first deep network model outputs a first probability value of the user to be cut over as a normal user and a second probability value of the user to be cut over as an abnormal user, respectively, based on the network configuration data characteristics of normal cutover users and network configuration data characteristics of abnormal cutover users.

[0045] The determining module is further configured to determine whether the user to be cutover is an abnormal cutover user based on the first probability value and the second probability value.

[0046] Thirdly, embodiments of the present invention provide an electronic device, comprising:

[0047] At least one processor; and

[0048] At least one memory communicatively connected to the processor, wherein:

[0049] The memory stores program instructions that can be executed by the processor, and the processor can execute the method described in the first aspect by calling the program instructions.

[0050] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions that cause the computer to perform the method described in the first aspect.

[0051] It should be understood that the second to fourth aspects of the embodiments of the present invention are consistent with the technical solutions of the first aspect of the embodiments of the present invention, and the beneficial effects achieved by each aspect and the corresponding feasible implementation are similar, and will not be described again. [Attached Image Description]

[0052] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 A flowchart of a method for determining abnormal cutover users provided in an embodiment of the present invention;

[0054] Figure 2 A flowchart of another method for determining abnormal cutover users provided in an embodiment of the present invention;

[0055] Figure 3 A flowchart of another method for determining abnormal cutover users provided in an embodiment of the present invention;

[0056] Figure 4 A flowchart of another method for determining abnormal cutover users provided in an embodiment of the present invention;

[0057] Figure 5 This is a schematic diagram of the structure of a first deep network model provided in an embodiment of the present invention;

[0058] Figure 6 A flowchart of another method for determining abnormal cutover users provided in an embodiment of the present invention;

[0059] Figure 7 This is a schematic diagram of the structure of an abnormal cutover user determination system provided in an embodiment of the present invention;

[0060] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

Detailed Implementation Methods

[0061] To better understand the technical solutions in this specification, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0062] It should be understood that the described embodiments are merely some, not all, of the embodiments in this specification. All other embodiments obtained by those skilled in the art based on the embodiments in this specification without inventive effort are within the scope of protection of this invention.

[0063] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of this specification. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0064] In this embodiment of the invention, the network configuration data of the user to be cut over is input into a first deep network model. The first deep network model determines whether the user to be cut over is an abnormal cutover user based on the characteristics of the network configuration data of normal cutover users and the characteristics of the network configuration data of abnormal cutover users.

[0065] Figure 1 This is a flowchart illustrating a method for determining abnormal cutover users according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0066] Step 101: Identify the users to be cut over from the users accessing the target cell. When the network occupancy rate of the target cell is high, network cutover can be performed on the PON ports with high occupancy rates. Specifically, based on actual needs, users corresponding to one or more optical splitters with high occupancy rates on a PON port can be migrated to PON ports with lower occupancy rates. The users corresponding to the one or more optical splitters are the aforementioned users to be cut over.

[0067] Step 102: Obtain the network configuration data of the user to be cut over. Specifically, the network configuration data may include the optical line terminal (OLT) equipment data (OLT device authentication code and PON port identification information, etc.) corresponding to the user to be cut over, the configuration data of the user to be cut over stored in the server (network account data, network authentication code, secondary optical splitter identification information, PON port identification information, and network room number information, etc.), and the actual wiring configuration data of the user to be cut over (network room number information, secondary optical splitter identification information, primary optical splitter identification information, and PON port identification information, etc.).

[0068] Step 103: Input the network configuration data into the first deep network model, so that the first deep network model outputs a first probability value for the user to be cut over as a normal cutover user and a second probability value for the user to be cut over as an abnormal cutover user, based on the network configuration data characteristics of normal cutover users and abnormal cutover users, respectively. The sum of the first probability value and the second probability value should be 100%.

[0069] In some embodiments, when inputting the network configuration data of the user to be cut over into the first deep network model, the network configuration data of the user to be cut over can be preprocessed. For example... Figure 2 As shown, the preprocessing procedure for the network configuration data of users to be cut over can include:

[0070] Step S1031: Based on the network configuration data, construct the network configuration dataset for the user to be cutover. This can be achieved by constructing a dataset of the OLT devices, a dataset stored on the server, and a dataset of the actual wiring configuration for the user. Optionally, only one dataset for the user to be cutover can be constructed, which includes the aforementioned OLT device data, the configuration data of the user to be cutover stored on the server, and the configuration data of the actual wiring for the user to be cutover.

[0071] Step S1032 involves performing a fixed difference operation or a one-hot encoding operation on the network configuration dataset. The fixed difference operation or one-hot encoding operation transforms the network configuration dataset from a 1×n-dimensional dataset into a two-dimensional network configuration data matrix, thus facilitating computation. Optionally, other methods can also be used to transform the network configuration dataset into a network configuration dataset matrix; no specific limitations are specified here.

[0072] Step S1033: Based on the calculation results, construct the network configuration data matrix of the user to be cut over.

[0073] Step S1034: Input the network configuration data matrix into the first deep network model so that the first deep network model can determine the network configuration data characteristics of the user to be cut over based on the network configuration data matrix, and output the first probability value of the user to be cut over as a normal user and the second probability value of the user to be cut over as an abnormal user based on the network configuration data characteristics of normal users and abnormal users.

[0074] Step 104: Determine whether the user to be cut over is an abnormal cutover user based on the first probability value and the second probability value. Optionally, if the second probability value is greater than 90%, it can be determined that the network configuration data of the user to be cut over is abnormal, and the user to be cut over is determined to be an abnormal cutover user.

[0075] In some embodiments, if the user to be cut over is an abnormal cutover user, the abnormal configuration parameters in the network configuration data of the user to be cut over can be identified, and the network configuration data of the abnormal cutover user can be repaired, such as... Figure 3 As shown, the processing steps of this method include:

[0076] Step 301: Input the network configuration data of the user to be cut over into the second deep network model, so that the second deep network model outputs the probability value of the user to be cut over as each category of abnormal problem based on the network configuration data characteristics of different categories of abnormal problems. Specifically, a network configuration data matrix can be constructed from the network configuration data according to the methods described in steps S1031 to S1034 above, and then the network configuration data matrix can be input into the second deep network model.

[0077] Step 302: Determine the anomaly category of the user to be cut over based on the probability values ​​of each anomaly category. Specifically, since the second deep network model actually outputs the probability that the network configuration data of the user to be cut over belongs to each anomaly category, the anomaly category with the highest probability can be determined as the anomaly category corresponding to the network configuration data of the user to be cut over. For example, if the probabilities of each anomaly category output by the second deep network model are: PON port anomaly (10%), first-level splitter anomaly (20%), and second-level splitter anomaly (70%), then the anomaly category of the user to be cut over can be determined as second-level splitter anomaly.

[0078] Step 303: Determine the abnormal configuration parameters of the user to be cut over based on the abnormal problem category. For example, if the abnormal problem category is secondary optical splitter abnormality, then the corresponding abnormal configuration parameters are the secondary optical splitter parameters.

[0079] Step 304: Based on the probability values ​​of the anomaly problems of the user to be cut over for each category output by the second deep network model, determine the anomaly problem category with the lowest probability value. The anomaly problem category with the lowest probability value can be selected from other anomaly problem categories associated with the determined anomaly problem category of the user to be cut over. For example, if the determined anomaly problem category of the user to be cut over is a secondary optical splitter anomaly, which is a type of wiring anomaly, other anomaly problem categories associated with the secondary optical splitter anomaly could be: primary optical splitter anomaly, PON port anomaly, etc.

[0080] Step 305: Determine the configuration parameter with the lowest anomaly probability based on the anomaly category with the lowest probability value.

[0081] Step 306: Determine the corrected value for the abnormal configuration parameter based on the value of the configuration parameter with the lowest probability of anomaly. For example, if the abnormal configuration parameter is a PON port, specifically PON port 5, and the parameter with the lowest probability of the corresponding abnormal problem category is a secondary optical splitter, specifically secondary optical splitter 1, then the correct primary optical splitter and PON port corresponding to secondary optical splitter 1 can be queried by traversing the server database. These will be secondary optical splitter 1 and PON port 1, respectively. Therefore, the correct value for the abnormal configuration parameter of the PON port for the user to be cut over is 5.

[0082] Step 307: Repair the abnormal configuration parameters according to the repair values.

[0083] In some embodiments, the training steps of the first deep network model described above are as follows: Figure 4 As shown:

[0084] Step 401: Obtain historical network configuration data, which includes network configuration data labeled with normal cutover user tags and network configuration data labeled with abnormal cutover user tags.

[0085] Step 402: Construct a historical network configuration data matrix based on the historical network configuration data. The method for constructing the historical network configuration data matrix is ​​the same as in steps S1031 to S1034, and will not be repeated here.

[0086] Step 403: Input the historical network configuration data matrix into the first deep network model for iterative training, so that the first deep network model learns the network configuration data characteristics of normal cutover users and abnormal cutover users. The iterative training process of the first deep network model is as follows: Figure 5 As shown, the first deep network model includes an input layer, at least one convolutional layer, at least one pooling layer, at least one fully connected layer, and an output layer. The input layer is the historical network configuration data matrix obtained above. Each neuron in the input layer is connected to the first hidden layer (convolutional kernel). The neurons in the first hidden layer (convolutional kernel) are only connected to neurons in the input layer in a local region; this local region is the local receptive field formed by the convolutional kernel. The convolutional kernel can be viewed as a window function. Therefore, the hidden layer (convolutional kernel) can acquire data features within the window function range. For example... Figure 5As shown, the window function range of the convolutional kernel is a 3×3 rectangle. By moving the convolutional kernel, it captures local features from different locations in the input layer and outputs them to the convolutional layer. Optionally, the stride and window size of the convolutional kernel can be adjusted according to actual needs. During the movement of the convolutional kernel, different data features at various locations in the historical network configuration data matrix of the input layer can be acquired. After at least one convolutional layer extracts data features from the historical network configuration data, the resulting data features are output to the pooling layer, which simplifies the data features output by the convolutional layer. Optionally, the pooling layer can also use a window function similar to the convolutional kernel when acquiring the data features output by the convolutional layer; this can be called a sampling layer. After simplifying the data features, the pooling layer outputs the result to the fully connected layer, where each neuron is fully connected to all neurons in the previous layer. The output of the last fully connected layer is output to the output layer. Optionally, the output layer can use softmax to classify the data features. In this way, after multiple iterations of training, the parameters in the first deep network model can be continuously updated. This allows the trained model to learn the differences in network configuration data between normal and abnormal cutover users.

[0087] Step 404: When the accuracy of the output result of the first deep network model is greater than the third threshold, the iteration ends. Optionally, the third threshold can be set to 90%.

[0088] In some embodiments, during iterative training of the second deep network model, since the second deep network model is used to identify the categories of anomalous problems, and each anomalous problem label corresponds to an anomalous configuration parameter, iterative training of the second deep network model using historical anomalous network configuration data labeled with different categories of anomalous problems allows the second deep network model to learn the data characteristics of network configuration data for different categories of anomalous problems. The training process is the same as in steps 401 to 404, and will not be repeated here.

[0089] In some embodiments, the presence of users requiring network cutover can be automatically identified based on the load information of the target cell, such as... Figure 6 As shown, the processing steps of this method include:

[0090] Step 601: Obtain the load information of each network interface of the target cell. The target cell contains multiple network interfaces, each network interface contains multiple optical splitters, and each optical splitter corresponds to multiple users. The load information includes multiple parameter items. The network interface can be a PON port, and the multiple parameter items in the load information can include: the number of 100Mbps network users, the number of Gigabit network users, bandwidth, and uplink bandwidth utilization for each network interface.

[0091] Step 602: Determine the score and weight of each parameter item for each network interface. In some embodiments, the entropy weight of each parameter item can be calculated first using information entropy, and then the weight of each parameter item can be calculated using the entropy weight method based on the entropy weight of each parameter item. In some embodiments, different scores can be set for different value ranges of each parameter item. For example, if the uplink bandwidth utilization rate is 83%, the score corresponding to the 80%-85% range is 80 points, then the score for the uplink bandwidth utilization rate is 80 points.

[0092] Step 603: Perform a weighted calculation on the scores and weights of each parameter item to obtain the load score for each network interface.

[0093] In some embodiments, subjective weights can be incorporated when calculating the load score for each network interface. Business experts, based on their experience, assign subjective weight values ​​to each parameter item. During weighted calculations, a combination of subjective and objective weights ensures that the final load score more accurately reflects the actual situation.

[0094] Step 604: Identify network interfaces with load scores greater than a first threshold as the network interfaces to be cut over. Preferably, the first threshold can be 80 points.

[0095] Step 605: Identify the optical splitters with a user count exceeding a second threshold from the optical splitters included in the network interface to be cut over as the optical splitters to be cut over. The number of network interfaces to be cut over can be determined based on load scores; the higher the score, the more network interfaces are available for cutover. For example, if PON port 1 has a load score of 80, then one optical splitter with more than 5 users can be selected from the multiple optical splitters corresponding to PON port 1 as the optical splitter to be cut over. If PON port 2 has a load score of 90, then three optical splitters with more than 5 users can be selected from the multiple optical splitters corresponding to PON port 2 as the optical splitters to be cut over.

[0096] Step 606: Determine the user corresponding to the splitter to be cut over as the user to be cut over.

[0097] Corresponding to the above-described method for determining abnormal cutover users, this embodiment of the invention provides a system for determining abnormal cutover users, such as... Figure 7 As shown, the system includes: a determination module 701, an acquisition module 702, and a deep learning module 703.

[0098] The determination module 701 is used to determine the users to be cut over from the users accessing the target cell.

[0099] Module 702 retrieves the network configuration data of the user to be cutoverdone.

[0100] The deep learning module 703 is used to input network configuration data into the first deep network model, so that the first deep network model outputs a first probability value of the user to be cut over as a normal user and a second probability value of the user to be cut over as an abnormal user, based on the network configuration data characteristics of normal cutover users and abnormal cutover users, respectively.

[0101] The determination module 701 is further configured to determine whether the user to be cut over is an abnormal cutover user based on the first probability value and the second probability value.

[0102] Figure 7 The abnormal cutover user determination system provided in the illustrated embodiment can be used to execute this specification. Figures 1-6 The implementation principle and technical effects of the method embodiment shown can be further referred to the relevant description in the method embodiment.

[0103] Figure 8 This is a schematic diagram of another electronic device provided in an embodiment of the present invention, such as... Figure 8 As shown, the electronic device described above may include at least one processor; and at least one memory communicatively connected to the processor, wherein the memory stores program instructions executable by the processor, and the processor can execute this specification by calling the program instructions. Figures 1-6 The embodiment shown provides a method for determining abnormal cutover users.

[0104] like Figure 8 As shown, the electronic device is represented in the form of a general-purpose computing device. The components of the electronic device may include, but are not limited to: one or more processors 810, communication interface 820 and memory 830, and a communication bus 840 connecting different system components (including memory 830, communication interface 820 and processing unit 810).

[0105] The communication bus 840 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0106] Electronic devices typically include a variety of computer-readable media. These media can be any available media that can be accessed by the electronic device, including volatile and non-volatile media, and removable and non-removable media.

[0107] Memory 830 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The electronic device may further include other removable / non-removable, volatile / non-volatile computer system storage media. Memory 830 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments described herein.

[0108] A program / utility having a set (at least one) of program modules may be stored in memory 830. Such program modules include—but are not limited to—an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. The program modules typically perform the functions and / or methods described in the embodiments of this specification.

[0109] Processor 810 executes various functional applications and data processing by running programs stored in memory 830, such as implementing the functions described in this specification. Figures 1-6 The embodiment shown provides a method for determining abnormal cutover users.

[0110] This specification provides a computer-readable storage medium that stores computer instructions that cause a computer to execute this specification. Figures 1-6 The embodiment shown provides a method for determining abnormal cutover users.

[0111] The aforementioned computer-readable storage medium may be any combination of one or more computer-readable media. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or flash memory, optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in connection with an instruction execution system, apparatus, or device.

[0112] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0113] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this specification. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0114] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this specification, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0115] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this specification includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which the embodiments of this specification pertain.

[0116] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0117] It should be noted that the devices involved in the embodiments of this specification may include, but are not limited to, personal computers (hereinafter referred to as PCs), personal digital assistants (hereinafter referred to as PDAs), wireless handheld devices, tablet computers, mobile phones, MP3 displays, MP4 displays, etc.

[0118] In the several embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0119] Furthermore, the functional units in the various embodiments of this specification can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.

[0120] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, a connector, or a network device, etc.) or a processor to execute some steps of the methods described in the various embodiments of this specification. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0121] The above description is merely a preferred embodiment of this specification and is not intended to limit this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of protection of this specification.

Claims

1. A method for determining users in abnormal cutovers, characterized in that, include: Identify the users to be cutovered from the users connected to the target cell; Obtain the network configuration data of the user to be cut over; The network configuration data is input into a first deep network model, so that the first deep network model outputs a first probability value that the user to be cut over is a normal user and a second probability value that is an abnormal user, respectively, based on the network configuration data characteristics of normal cutover users and network configuration data characteristics of abnormal cutover users. Based on the first probability value and the second probability value, determine whether the user to be cut over is an abnormal cutover user; The method further includes: If the user to be cut over is an abnormal user, then the network configuration data of the user to be cut over is input into the second deep network model, so that the second deep network model outputs the probability value of the user to be cut over as an abnormal problem of each category according to the network configuration data characteristics of different categories of abnormal problems. The anomaly category of the user to be cut over is determined based on the probability value of each category of anomaly. Based on the category of the abnormal problem, determine the abnormal configuration parameters of the user to be cut over; Based on the probability values ​​of the user to be cut over as an abnormal problem of each category output by the second deep network model, determine the abnormal problem category with the smallest probability value; Based on the anomaly category with the lowest probability value, determine the configuration parameter with the lowest anomaly probability; The repair value of the abnormal configuration parameter is determined based on the value of the configuration parameter with the lowest probability of abnormality. The abnormal configuration parameters are repaired based on the repair values.

2. The method according to claim 1, characterized in that, Identify the users to be cutoverdone from the users accessing the target cell, including: Obtain the load information of each network interface of the target cell, wherein the target cell contains multiple network interfaces, each network interface contains multiple optical splitters, each optical splitter corresponds to multiple users, and the load information contains multiple parameter items; Determine the score and weight of each parameter item for each network interface; The scores and weights of each parameter item are weighted to obtain the load score for each network interface; The network interfaces whose load scores are greater than the first threshold are identified as network interfaces to be cut over. The optical splitter containing more than a second threshold number of users is identified from the optical splitters included in the network interface to be cut over as the optical splitter to be cut over. The user corresponding to the splitter to be cut is identified as the user to be cut.

3. The method according to claim 1, characterized in that, The network configuration data is input into a first deep network model, so that the first deep network model outputs a first probability value for the user to be cut over as a normal user and a second probability value for the user to be cut over as an abnormal user, based on the network configuration data characteristics of normal cutover users and abnormal cutover users, respectively, including: Based on the network configuration data, construct the network configuration dataset of the users to be cutoverdated; Perform fixed difference operation or one-hot encoding operation on the network configuration dataset; Based on the calculation results, construct the network configuration data matrix for the users to be cut over; The network configuration data matrix is ​​input into the first deep network model so that the first deep network model determines the network configuration data characteristics of the user to be cut over based on the network configuration data matrix, and outputs a first probability value that the user to be cut over is a normal user and a second probability value that is an abnormal user based on the network configuration data characteristics of normal users and abnormal users.

4. The method according to claim 1, characterized in that, The training process of the first deep network model includes: Obtain historical network configuration data, which includes network configuration data labeled with normal cutover users and network configuration data labeled with abnormal cutover users; Construct a historical network configuration data matrix based on the historical network configuration data; The historical network configuration data matrix is ​​input into the first deep network model for iterative training, so that the first deep network model learns the network configuration data characteristics of normal cutover users and the network configuration data characteristics of abnormal cutover users. The iteration ends when the accuracy of the output result of the first deep network model is greater than the third threshold.

5. The method according to claim 1, characterized in that, The training process of the second deep network model includes: Obtain historical abnormal network configuration data, which includes abnormal problem tags labeled with different categories, and each abnormal problem tag corresponds to an abnormal configuration parameter; Construct a historical abnormal network configuration data matrix based on the historical abnormal network configuration data; The historical anomaly network configuration data matrix is ​​input into the second deep network model for iterative training, so that the second deep network model learns the network configuration data features corresponding to each category of anomalies. The iteration ends when the accuracy of the output result of the second deep network model is greater than the fourth threshold.

6. A system for determining users during abnormal cutovers, characterized in that, include: The determination module is used to identify users to be cut over from the users accessing the target cell; The acquisition module acquires the network configuration data of the user to be cutoverdone. The deep learning module is used to input the network configuration data into the first deep network model, so that the first deep network model outputs a first probability value of the user to be cut over as a normal user and a second probability value of the user to be cut over as an abnormal user, respectively, based on the network configuration data characteristics of normal cutover users and network configuration data characteristics of abnormal cutover users. The determining module is further configured to determine whether the user to be cutover is an abnormal cutover user based on the first probability value and the second probability value. The deep learning module is also used to input the network configuration data of the user to be cut over into the second deep network model when the user to be cut over is an abnormal user, so that the second deep network model outputs the probability value of the user to be cut over as an abnormal problem of each category according to the network configuration data characteristics of different categories of abnormal problems. The determining module is further configured to determine the abnormal problem category of the user to be cut over based on the probability value of each category of abnormal problem, and to determine the abnormal configuration parameters of the user to be cut over based on the abnormal problem category. The determining module is further configured to determine the abnormal problem category with the smallest probability value based on the probability values ​​of the user to be cut over as each category of abnormal problem output by the second deep network model; and to determine the configuration parameter with the smallest abnormal probability based on the abnormal problem category with the smallest probability value. Based on the value of the configuration parameter with the lowest probability of anomaly, determine the repair value of the abnormal configuration parameter; and repair the abnormal configuration parameter based on the repair value.

7. An electronic device, characterized in that, include: At least one processor; as well as At least one memory communicatively connected to the processor, wherein: The memory stores program instructions that can be executed by the processor, and the processor can execute the method as described in any one of claims 1 to 5 by calling the program instructions.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause the computer to perform the method as described in any one of claims 1 to 5.

Images