Data-driven optical cat network terminal data anomaly monitoring system

Abstract

The application discloses a data-driven optical modem network terminal data anomaly monitoring system, which comprises a terminal data acquisition module, an associated network construction module, a terminal associated similarity construction module, a core terminal discrimination module, a terminal service group division module, an isolated optical modem terminal attribution determination module and a terminal data anomaly monitoring module. The application belongs to the field of data processing and specifically refers to a data-driven optical modem network terminal data anomaly monitoring system. The present scheme divides two types of scenes with or without a public associated transit terminal, calculates terminal associated similarity respectively, introduces a running state decay factor to filter invalid connection interference, compares terminal real-time time delay and packet loss rate with the average benchmark of the whole network, dynamically corrects link associated weight, constructs a service type traction index, accumulates the traction strength according to the service flow weight of the marked terminal in the surrounding, matches the attribution type according to the maximum traction index, and further improves the optical modem network terminal data anomaly monitoring effect.

Description

Technical Field

[0001] This invention relates to the field of data processing, specifically to a data-driven optical modem network terminal data anomaly monitoring system. Background Technology

[0002] The optical modem network terminal data anomaly monitoring system is a professional operation and maintenance monitoring system designed for all optical modem terminals across the operator's network. It relies on operational data such as network topology, service traffic, latency, and packet loss to automatically identify terminal offline status, link degradation, traffic anomalies, clustered faults, and abnormal correlation behaviors. This enables the automatic discovery, location, and early warning of potential optical modem faults across the entire network. However, general optical modem network terminal data anomaly monitoring systems often suffer from difficulties in accurately identifying the spread of multi-terminal clustered abnormal behaviors, are susceptible to interference and misjudgments from botnet connections, and thus have poor accuracy in terminal data anomaly monitoring. Furthermore, these systems typically have monitoring blind spots, and cannot horizontally compare the aggregation characteristics of optical modem terminals with different network sizes, leading to poor terminal data anomaly monitoring results. Summary of the Invention

[0003] To address the above issues and overcome the shortcomings of existing technologies, this invention provides a data-driven optical modem network terminal data anomaly monitoring system. Addressing the problems of general optical modem network terminal data anomaly monitoring systems, such as difficulty in accurately identifying multi-terminal group abnormal linkage and diffusion behaviors, susceptibility to botnet interference and misjudgments, and consequently poor accuracy in terminal data anomaly monitoring, this solution divides the scenarios into two categories: those with and without shared relay terminals. It calculates terminal association similarity separately for each scenario, supporting accurate determination of multi-terminal overlapping abnormal associations and group linkage diffusion. It introduces an operating state attenuation factor based on latency and packet loss rate to penalize botnet links that exist in the topology but have extremely poor communication quality, filtering out invalid connection interference. By comparing the terminal's real-time latency and packet loss rate with the network-wide average benchmark, and controllably adjusting the state difference penalty intensity through a sensitivity factor, it dynamically corrects the link association weights to adapt to the terminal's operating state. Dynamically changing; addressing the issues of blind spots in general optical modem network terminal anomaly monitoring systems and the inability to horizontally compare the aggregation characteristics of optical modem terminals of different network sizes, leading to poor terminal data anomaly monitoring results, this solution defines an optical modem terminal aggregation factor. This factor quantifies the network value of terminals from dimensions such as service aggregation degree and topology representativeness. Core terminals are selected using the network-wide average aggregation factor, providing a reliable benchmark for subsequent group division and anomaly monitoring. For non-isolated terminals, aggregation factors are constructed to objectively characterize the saturation and regularity of service aggregation in local subnets, enabling horizontal comparison of the aggregation degree of terminals in different locations and of different sizes. A service type traction index is constructed, calculating the traction strength based on the cumulative weight of surrounding marked terminal service traffic. The maximum traction index is used to match the attribution type, and the attribution of isolated terminals is automatically determined by the actual service link weight. This ultimately improves the effectiveness of optical modem network terminal data anomaly monitoring.

[0004] The technical solution adopted by the present invention is as follows: The data-driven optical modem network terminal data anomaly monitoring system provided by the present invention includes a terminal data acquisition module, an associated network construction module, a terminal association similarity construction module, a core terminal identification module, a terminal service group division module, an isolated optical modem terminal affiliation determination module, and a terminal data anomaly monitoring module;

[0005] The terminal data acquisition module acquires the operating data of all optical modem terminals in the network, and after standardized preprocessing, it forms a monitoring dataset.

[0006] The associated network construction module constructs a network of associated optical modem terminals across the entire network based on the monitoring dataset.

[0007] The terminal association similarity construction module defines the optical modem terminal aggregation factor based on the whole network optical modem terminal association network, divides isolated optical modem terminals and non-isolated optical modem terminals, and constructs terminal association similarity for each.

[0008] The core terminal identification module defines the optical modem terminal aggregation factor to identify the core terminal.

[0009] The terminal service group division module uses the identified core terminals as the initial queue for diffusion and completes the terminal group labeling and division based on terminal association similarity.

[0010] The isolated optical modem terminal attribution determination module constructs a service type influence score for isolated optical modem terminals that have not completed group labeling, thereby realizing terminal attribution determination;

[0011] The terminal data anomaly monitoring module monitors terminal data anomalies by comparing the deviation of a single terminal from the overall group based on the terminal group division results.

[0012] Furthermore, the association network construction module constructs a network of optical modem terminals across the entire network based on the monitoring dataset: Let V be the non-empty finite set of optical modem terminals across the entire network, and let the node unit be the optical modem terminal. If there is a direct service interaction between two terminals, then the connection edge is formed. The service traffic association weight between any two directly connected optical modem terminals is set to the normalized terminal Euclidean distance. When there is no direct service interaction, the service traffic association weight is 0.

[0013] Furthermore, the terminal association similarity construction module is based on the entire network of optical modem terminal associations, and objectively calculates the optical modem terminal association similarity index by combining the network topology and edge weights; it distinguishes between two natural topology scenarios with and without public associated relay terminals, and quantifies the degree of service association between two optical modem terminals; at the same time, it introduces an operating status attenuation factor, and on the basis of topology connection, it penalizes those links that have connections but whose communication quality is severely degraded.

[0014] Furthermore, the core terminal identification module defines an optical modem terminal aggregation factor based on the network of optical modem terminals associated with the entire network. This factor is used to quantitatively evaluate the degree of service aggregation and topological representativeness of each terminal in the network. Only terminals with aggregation factors not lower than the average level of the entire network are identified as core terminals.

[0015] Furthermore, the terminal service group division module automatically expands hierarchically from the core terminal, traversing the surrounding optical modem terminals, and objectively divides the terminals across the entire network into service groups based on service association similarity; only when the association similarity between terminals is greater than or equal to a preset monitoring threshold, it is determined that two terminals belong to the same linked service group.

[0016] Furthermore, the isolated optical modem terminal attribution determination module automatically calculates the traction index of different service types on the legacy terminal based on the actual service link weight of the surrounding terminals that have completed group marking, and objectively matches the associated service type according to the maximum traction index for isolated optical modem terminals that have not completed group marking.

[0017] Furthermore, the terminal data anomaly monitoring module, based on the merged business linkage groups, calculates the deviation of a single terminal's operating indicators from the overall steady-state distribution of the group. It adaptively defines the group's normal steady-state distribution range using the statistical distribution characteristics of terminal operating indicators within the group, and determines the upper and lower limits of the steady-state range for each dimension of indicators using a mean-standard deviation statistical method. It then calculates the deviation of a single terminal's operating indicators from the overall steady-state distribution of the group. When multiple dimensions of a terminal's operating indicators exceed the normal steady-state distribution range of the corresponding indicators in the group, it determines that the optical modem terminal has an operating anomaly. Similarly, when multiple types of business terminals meet the deviation of their operating indicators from the overall steady-state distribution of the group, it determines that the optical modem terminal has an operating anomaly.

[0018] The beneficial effects achieved by adopting the above solution are as follows:

[0019] (1) In view of the problem that the general optical modem network terminal data anomaly monitoring system has difficulty in accurately identifying the abnormal linkage and diffusion behavior of multiple terminals, and is easily misjudged by zombie connection interference, which leads to poor accuracy of terminal data anomaly monitoring, this solution divides the scenarios into two categories: those with and without common associated relay terminals, calculates the terminal association similarity separately, and supports accurate judgment of multiple terminal overlapping abnormal association and group linkage diffusion; introduces an operation status attenuation factor based on latency and packet loss rate, applies weight penalty to zombie links that exist in the topology but have extremely poor communication quality, and filters out invalid connection interference; compares the terminal real-time latency and packet loss rate with the network average benchmark, and controls the state difference penalty intensity through sensitivity factor to dynamically correct the link association weight and adapt to the dynamic changes in terminal operation status.

[0020] (2) To address the problem that general optical modem network terminal data anomaly monitoring systems have monitoring blind spots and cannot horizontally compare the aggregation characteristics of optical modem terminals of different network sizes, resulting in poor terminal data anomaly monitoring effects, this solution defines an optical modem terminal aggregation factor to quantify the terminal network value from the dimensions of service aggregation degree and topology representativeness. Core terminals are selected using the average aggregation factor of the entire network, providing a reliable benchmark for subsequent group division and anomaly monitoring. For non-isolated terminals, an aggregation factor is constructed to objectively characterize the saturation and regularity of service aggregation in local subnets, enabling horizontal comparison of the aggregation degree of terminals in different locations and of different sizes. A service type traction index is constructed, and the traction strength is calculated by accumulating the service traffic weights of surrounding marked terminals. The type of ownership is matched according to the maximum traction index, and the ownership of isolated terminals is automatically divided by the weight of the actual service link. This improves the effectiveness of optical modem network terminal data anomaly monitoring. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the data anomaly monitoring system for optical modem network terminals based on data driving provided by the present invention.

[0022] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0023] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0024] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0025] Example 1, see Figure 1 The data-driven optical modem network terminal data anomaly monitoring system provided by the present invention includes a terminal data acquisition module, an associated network construction module, a terminal association similarity construction module, a core terminal identification module, a terminal service group division module, an isolated optical modem terminal affiliation determination module, and a terminal data anomaly monitoring module.

[0026] The terminal data acquisition module acquires the operating data of all optical modem terminals in the network, and after standardized preprocessing, it forms a monitoring dataset.

[0027] The associated network construction module constructs a network of associated optical modem terminals across the entire network based on the monitoring dataset.

[0028] The terminal association similarity construction module defines the optical modem terminal aggregation factor based on the whole network optical modem terminal association network, divides isolated optical modem terminals and non-isolated optical modem terminals, and constructs terminal association similarity for each.

[0029] The core terminal identification module defines the optical modem terminal aggregation factor to identify the core terminal.

[0030] The terminal service group division module uses the identified core terminals as the initial queue for diffusion and completes the terminal group labeling and division based on terminal association similarity.

[0031] The isolated optical modem terminal attribution determination module constructs a service type influence score for isolated optical modem terminals that have not completed group labeling, thereby realizing terminal attribution determination;

[0032] The terminal data anomaly monitoring module monitors terminal data anomalies by comparing the deviation of a single terminal from the overall group based on the terminal group division results.

[0033] Example 2, see Figure 1 This embodiment is based on the above embodiment. The terminal data acquisition module acquires the operating data of the optical modem terminals across the entire network and preprocesses the acquired data. The standardized preprocessing is completed through data cleaning, maximum and minimum normalization, and outlier removal to form a monitoring dataset. The operating data of the optical modem terminals across the entire network includes uplink and downlink bandwidth, network latency, packet loss rate, online dwell time, port traffic, number of connection sessions, and signal attenuation value.

[0034] Example 3, see Figure 1 This embodiment is based on the above embodiment. The association network construction module constructs a network of optical modem terminals based on the monitoring dataset: Let V be a non-empty finite set of optical modem terminals in the whole network, and let the node unit be an optical modem terminal. If there is a direct service interaction between two terminals, then the connection edge is formed. The service traffic association weight W(u,v) between any two directly connected optical modem terminals u and v is set to the normalized terminal Euclidean distance. When there is no direct service interaction, the service traffic association weight is 0.

[0035] Example 4, see Figure 1This embodiment is based on the above embodiment. The terminal association similarity construction module is based on the entire network of optical modem terminal associations. It objectively calculates the optical modem terminal association similarity index by combining the network topology and edge weights. It distinguishes between two natural topology scenarios with and without public association relay terminals, quantifies the degree of service association between two optical modem terminals, and uses it as an objective basis for judging abnormal linkage and spread. It is adapted to traffic networks and supports the judgment of abnormal associations with multiple overlapping terminals. At the same time, in order to avoid zombie connections (connections with extremely poor quality) being misjudged as strong associations, a running status decay factor is introduced. On the basis of topology connection, it punishes those links that have connections but severely degraded communication quality, so that the similarity index is more in line with the nature of healthy service associations.

[0036] A state consistency decay factor is constructed based on network latency and packet loss rate, and is expressed as follows: ;in, and It refers to the terminal's real-time network latency and packet loss rate; and It represents the average latency and packet loss rate across the entire network; It is a sensitivity factor, with a value of 1 to 5, which controls the degree of penalty imposed on the weight by the difference in the control state;

[0037] When there are no shared intermediate terminals, the terminal association similarity is represented as follows: ;

[0038] When there are shared transit terminals, the terminal association similarity is represented as follows: ;

[0039] in, ; ;

[0040] in, This refers to the terminal association similarity; the higher the value, the stronger the association. It is the set of common associated relay terminals of terminals u and v; It is the number of internally connected edges; It refers to the number of terminals; and These are the total correlation between terminal u and terminal v, representing the total number of overall service connections for the terminals; x is the public relay terminal. It is the sum of path association weights from terminal u to transit terminal x to v; It is the common correlation strength factor between u and v; It takes the minimum value of the sum of the correlation between terminals u and v for standardization; all calculations are automatically generated by network topology and traffic weights.

[0041] By performing the above operations, this solution addresses the problem that general optical modem network terminal data anomaly monitoring systems often struggle to accurately identify multi-terminal group abnormal linkage and diffusion behaviors, are susceptible to botnet interference and misjudgments, and consequently suffer from poor accuracy in terminal data anomaly monitoring. This solution categorizes scenarios into two types: those with and without shared relay terminals, calculating terminal association similarity separately to support accurate determination of multi-terminal overlapping abnormal associations and group linkage diffusion. It introduces an operational state attenuation factor based on latency and packet loss rate to penalize botnet links that exist in the topology but have extremely poor communication quality, filtering out invalid connection interference. Furthermore, by comparing the terminal's real-time latency and packet loss rate with the network-wide average benchmark, and controlling the state difference penalty intensity through a sensitivity factor, the link association weight is dynamically corrected to adapt to dynamic changes in terminal operational state.

[0042] Example 5, see Figure 1 This embodiment is based on the above embodiment. The core terminal identification module is based on the network of optical modem terminals associated with the entire network. According to the natural characteristics of network service aggregation, it defines the optical modem terminal aggregation factor to quantitatively evaluate the degree of service aggregation and topological representativeness of each terminal in the network. Only terminals with aggregation factors not lower than the average level of the entire network are identified as core terminals to ensure that the benchmark terminal has typical service characteristics and stable association characteristics.

[0043] Isolated optical modem terminal determination rule: When the number of adjacent terminals of terminal u or When the terminal is identified as an isolated optical modem terminal, the corresponding terminal aggregation factor is 1.

[0044] The aggregation factor of a non-isolated optical modem terminal is expressed as: ; ;

[0045] molecular Objectively reflects the total service traffic intensity of the terminal's adjacent subnetwork, naturally determined by actual service interactions; denominator It is the maximum possible number of edges connecting adjacent terminals, which is naturally determined by the network topology. It is the maximum value of the adjacent edge weight, used to normalize the aggregation factor to the [0,1] interval. The normalization process is determined by the weight distribution of the network itself. The overall factor represents the saturation and regularity of the aggregation of local sub-network services in the terminal, which conforms to the natural law of the distribution of complex network services.

[0046] The average aggregation factor of all optical modems in the network is calculated and expressed as: ;

[0047] The core terminal identification rules are expressed as follows: Terminals that meet the criteria are core terminals with high business concentration, dense adjacency relationships, and strong regional representativeness.

[0048] in, It is the aggregation factor of terminal u; x and y are any two different adjacent terminals; It is the service traffic association weight between adjacent terminals x and y; It is the maximum service traffic association weight between adjacent terminals of terminal u, used for factor normalization; This represents the total number of optical modem terminals across the entire network. It is the average aggregation factor across the entire network.

[0049] Example 6, see Figure 1 This embodiment is based on the above embodiment. The terminal service group division module takes advantage of the inherent hierarchical propagation characteristics of network faults, and automatically spreads hierarchically from the core terminal to the surrounding optical modem terminals. Based on the service association similarity, it objectively divides the terminals in the entire network into service groups. Only when the association similarity between terminals is greater than or equal to the preset monitoring threshold, it is determined that the service links of the two terminals are highly coupled and belong to the same linkage service group, ensuring that the divided groups have real service connectivity association. Terminals that cannot be directly associated with a group are temporarily stored and processed centrally later to avoid monitoring blind spots.

[0050] The process logic is as follows:

[0051] (1) Initialize the queue, send the identified core terminals into the monitoring sequence, and complete the group initialization;

[0052] (2) Traverse, take out one terminal from the sequence, and traverse all adjacent optical modem terminals;

[0053] (3) Determine the association within the same group, if the following conditions are met. If the terminal has not completed group labeling, it is determined that the two belong to the same business linkage group, the same group attribute is labeled and sent into the sequence to continue hierarchical diffusion; This is the threshold for association similarity monitoring, with a value ranging from 0.3 to 0.7;

[0054] (4) When the sequence traversal is empty, complete the delineation of a group of business linkage terminals;

[0055] (5) Isolated optical modem terminals that have not completed group marking have weak link coupling and cannot be included in the cluster. They will be transferred to the subsequent unified judgment process for legacy terminals.

[0056] Each optical modem terminal group corresponds to a set of service types.

[0057] Example 7, see Figure 1This embodiment is based on the above embodiment. The isolated optical modem terminal affiliation determination module is for isolated optical modem terminals that have not completed group marking. Because their own connections are sparse, they cannot be directly determined through diffusion at the base terminal level, which can easily form a monitoring blind spot. Therefore, based on the real service link weight of the surrounding terminals that have completed the same group marking, it automatically calculates the traction index of different service types on the legacy terminals, and objectively matches the service association type according to the maximum traction index. When multiple types have the same score, the terminal is allowed to belong to different service association types, which is suitable for the actual networking scenario of multiple services overlapping and coupled. By setting two levels of adaptive thresholds, the association type identification is completed, and finally the service group affiliation merging of all optical modem terminals in the network is achieved without omission and with full coverage.

[0058] The traction index of isolated optical modem terminal u corresponding to service type l Represented as: ;

[0059] Judgment rules: Select the service type with the highest traction index as the type to which the legacy terminal belongs; if the scores of multiple service types are equal or within the adaptive threshold range, it is determined to be a multiple service superposition association type;

[0060] The specific settings of the hyperparameters in the scheme are obtained based on offline data fitting and traversal optimization.

[0061] By performing the above operations, this solution addresses the issues of blind spots in general optical modem network terminal anomaly monitoring systems, where optical modem terminals of different network sizes cannot be horizontally compared in terms of aggregation characteristics, leading to poor terminal data anomaly monitoring results. This solution defines an aggregation factor for optical modem terminals, quantifying the network value of terminals from the dimensions of service aggregation degree and topological representativeness. Core terminals are selected using the network-wide average aggregation factor, providing a reliable benchmark for subsequent group division and anomaly monitoring. For non-isolated terminals, an aggregation factor is constructed to objectively characterize the saturation and regularity of service aggregation in local subnets, enabling horizontal comparison of aggregation degrees for terminals of different locations and sizes. A service type traction index is constructed, calculating traction strength based on the cumulative weight of surrounding marked terminal service traffic, matching the attribution type according to the maximum traction index, and automatically assigning isolated terminals based on the actual service link weight. This ultimately improves the effectiveness of optical modem network terminal data anomaly monitoring.

[0062] Example 8, see Figure 1This embodiment is based on the above embodiment. The terminal data anomaly monitoring module calculates the deviation of the operating indicators of a single terminal from the overall steady-state distribution of the group based on the merged business linkage group. It adaptively defines the normal steady-state distribution range of the group by using the statistical distribution characteristics of the terminal operating indicators within the group. It uses the mean standard deviation statistical method to determine the upper and lower limits (3σ) of the steady-state range of each dimension indicator. It calculates the deviation of the operating indicators of a single terminal from the overall steady-state distribution of the group. When the operating indicators of a terminal in multiple dimensions exceed the normal steady-state distribution range of the corresponding indicators in the group, it is determined that the optical modem terminal has an operating anomaly. When multiple types of business terminals meet the deviation of their operating indicators from the overall steady-state distribution of the group, it is determined that the optical modem terminal has an operating anomaly.

[0063] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0064] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A data-driven optical modem network terminal data anomaly monitoring system, characterized in that: The system includes a terminal data acquisition module, an association network construction module, a terminal association similarity construction module, a core terminal identification module, a terminal service group division module, an isolated optical modem terminal attribution determination module, and a terminal data anomaly monitoring module; The terminal data acquisition module acquires the operating data of all optical modem terminals in the network, and after standardized preprocessing, it forms a monitoring dataset. The associated network construction module constructs a network of associated optical modem terminals across the entire network based on the monitoring dataset. The terminal association similarity construction module defines the optical modem terminal aggregation factor based on the whole network optical modem terminal association network, divides isolated optical modem terminals and non-isolated optical modem terminals, and constructs terminal association similarity for each. The core terminal identification module defines the optical modem terminal aggregation factor to identify the core terminal. The terminal service group division module uses the identified core terminals as the initial queue for diffusion and completes the terminal group labeling and division based on terminal association similarity. The isolated optical modem terminal attribution determination module constructs a service type influence score for isolated optical modem terminals that have not completed group labeling, thereby realizing terminal attribution determination; The terminal data anomaly monitoring module monitors terminal data anomalies by comparing the deviation of a single terminal from the overall group based on the terminal group division results.

2. The data-driven optical modem network terminal data anomaly monitoring system according to claim 1, characterized in that: The terminal association similarity construction module is based on the network of optical modem terminal associations across the entire network. It objectively calculates the optical modem terminal association similarity index by combining the network topology and edge weights. It distinguishes between two natural topology scenarios with and without public associated relay terminals to quantify the degree of service association between two optical modem terminals. At the same time, it introduces an operating status attenuation factor to penalize links that exist but have severely degraded communication quality, based on topology connections.

3. The data-driven optical modem network terminal data anomaly monitoring system according to claim 2, characterized in that: The core terminal identification module defines an optical modem terminal aggregation factor based on the network of optical modem terminals associated with the entire network. This factor is used to quantitatively evaluate the degree of service aggregation and topological representativeness of each terminal in the network. Only terminals with an aggregation factor not lower than the average level of the entire network are identified as core terminals.

4. The data-driven optical modem network terminal data anomaly monitoring system according to claim 3, characterized in that: The terminal service group division module automatically expands hierarchically from the core terminal, traversing the surrounding optical modem terminals, and objectively divides the terminals across the entire network into service groups based on service association similarity; only when the association similarity between terminals is greater than or equal to the preset monitoring threshold, it is determined that two terminals belong to the same linkage service group.

5. The data-driven optical modem network terminal data anomaly monitoring system according to claim 4, characterized in that: The isolated optical modem terminal attribution determination module automatically calculates the traction index of different service types on the legacy terminal based on the actual service link weight of the surrounding terminals that have completed group marking, and objectively matches the associated service type according to the maximum traction index for isolated optical modem terminals that have not completed group marking.

6. The data-driven optical modem network terminal data anomaly monitoring system according to claim 5, characterized in that: The terminal data anomaly monitoring module, based on the merged business linkage groups, calculates the deviation of a single terminal's operating indicators from the overall steady-state distribution of the group. It adaptively defines the group's normal steady-state distribution range using the statistical distribution characteristics of terminal operating indicators within the group, and determines the upper and lower limits of the steady-state range for each dimension of indicators using a mean-standard deviation statistical method. When a terminal's multi-dimensional operating indicators exceed the normal steady-state distribution range of the corresponding indicators in the group, the optical modem terminal is determined to have an operational anomaly. Similarly, when multiple types of business terminals meet the deviation of their operating indicators from the overall steady-state distribution of the group, the optical modem terminal is also determined to have an operational anomaly.

7. The data-driven optical modem network terminal data anomaly monitoring system according to claim 6, characterized in that: The associated network construction module constructs a network of optical modem terminals across the entire network based on the monitoring dataset: Let V be a non-empty finite set of optical modem terminals across the entire network, and let the node unit be an optical modem terminal. If there is a direct service interaction between two terminals, then the connection edge is formed. The service traffic association weight between any two directly connected optical modem terminals is set to the normalized terminal Euclidean distance. When there is no direct service interaction, the service traffic association weight is 0.

Images