A device data acquisition method, electronic device, and program product

Abstract

Embodiments of the present disclosure provide a device data acquisition method, an electronic device and a program product; the device data acquisition method comprises: acquiring a device association model corresponding to a network, the device association model being used to represent connection strength between a plurality of communication devices in the network and device states of the communication devices; performing multi-dimensional reasoning on the communication devices based on the device association model to obtain a multi-dimensional reasoning result; and performing relationship extraction on the plurality of communication devices based on the device association model and the multi-dimensional reasoning result to obtain a device knowledge graph in the network. Through the embodiment scheme, complex relationship changes between devices can be effectively coped with, data with depth and actual guiding significance can be extracted, a clear and compact data basis is provided for subsequent reasoning, and accurate expression and efficient storage of event data are ensured.

Description

Technical Field

[0001] This disclosure relates to the field of communication networks, and in particular to a device data acquisition method, electronic device, and program product. Background Technology

[0002] Based on the characteristics of communication failure data, the industry faces the following pain points when extracting triples from knowledge graphs:

[0003] 1) Lack of effective data modeling mechanisms: Most triple extraction methods for related fault data analysis rely on an explicit perspective and fail to effectively integrate the connection relationships between devices, fault states, etc. For the extraction of complex fault propagation and causal chains, the relevant methods lack flexible data modeling mechanisms and are unable to effectively cope with the complex changes in relationships between devices.

[0004] 2) Difficulty in mining non-explicit knowledge: Communication fault data contains a large amount of non-explicit information. For example, the coherence information between devices often lacks clear labels or explicit features. Related triple extraction methods usually struggle to effectively infer meaningful knowledge from this explicit information, resulting in extracted knowledge lacking depth and practical guiding significance. Summary of the Invention

[0005] This disclosure provides a device data acquisition method, an electronic device, and a program product.

[0006] In a first aspect, embodiments of this disclosure provide a device data acquisition method, which includes:

[0007] Obtain the network-corresponding device association model, which is used to represent the connection strength between multiple communication devices in the network and the device status of each communication device;

[0008] Based on the device association model, multidimensional reasoning is performed on the communication device to obtain multidimensional reasoning results;

[0009] Based on the device association model and the multidimensional reasoning results, the relationships of the multiple communication devices are extracted to obtain the device knowledge graph in the network.

[0010] Secondly, embodiments of this disclosure also provide an electronic device, including:

[0011] One or more processors;

[0012] A memory having stored one or more programs, which, when executed by one or more processors, enable the one or more processors to implement the device data acquisition method;

[0013] One or more input / output (I / O) interfaces are connected between the processor and the memory and configured to enable information exchange between the processor and the memory.

[0014] Thirdly, this disclosure also provides a computer program product, which includes a computer program that, when executed by a processor, implements the device data acquisition method.

[0015] This embodiment of the scheme obtains a network-corresponding device association model, which represents the connection strength between multiple communication devices in the network and the device status of each communication device. Based on the device association model, multi-dimensional reasoning is performed on the communication devices to obtain multi-dimensional reasoning results. Based on the device association model and the multi-dimensional reasoning results, relationships are extracted from multiple communication devices to obtain a device knowledge graph in the network. Through this embodiment, multi-dimensional reasoning based on the device association model can accurately capture the complex state relationships between devices in the communication network, overcoming the shortcomings of traditional methods in dealing with state propagation and causal chain analysis. It can effectively cope with complex relationship changes between devices, extract data with depth and practical guiding significance, provide a clear and compact data foundation for subsequent reasoning, and ensure the accurate expression and efficient storage of event data. Attached Figure Description

[0016] In the accompanying drawings of the embodiments disclosed herein:

[0017] Figure 1 A flowchart of a device data acquisition method provided in this embodiment of the disclosure;

[0018] Figure 2 This is a schematic diagram of a device data acquisition method provided in an embodiment of the present disclosure;

[0019] Figure 3 This is a flowchart of a method for obtaining a network-corresponding device association model provided in an embodiment of the present disclosure;

[0020] Figure 4 A flowchart of a method for dynamically inferring the strength of causal relationships between multiple communication devices based on a device association model and outputting causal inference results, provided in an embodiment of this disclosure;

[0021] Figure 5 A flowchart of a method for determining the probability of a device state propagating from one communication device to another based on a device association model and causal inference results, and outputting the propagation inference results, is provided in this embodiment of the disclosure.

[0022] Figure 6 This is a block diagram of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation

[0023] To enable those skilled in the art to better understand the technical solutions of this disclosure, the communication-sensing data processing method and computer-readable storage medium provided in the embodiments of this disclosure will be described in detail below with reference to the accompanying drawings.

[0024] The present disclosure will be described more fully below with reference to the accompanying drawings; however, the embodiments shown may be embodied in different forms, and the present disclosure should not be construed as limited to the embodiments set forth below. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will enable those skilled in the art to fully understand the scope of the disclosure.

[0025] The accompanying drawings of the embodiments disclosed herein are provided to further illustrate the embodiments of this disclosure and form part of the specification. They are used together with the detailed embodiments to explain this disclosure and do not constitute a limitation thereof. The above and other features and advantages will become more apparent to those skilled in the art from the description of the detailed embodiments with reference to the accompanying drawings.

[0026] This disclosure may be described with reference to plan and / or cross-sectional views using the ideal schematic diagrams of this disclosure. Therefore, the example illustrations may be modified according to manufacturing techniques and / or tolerances.

[0027] Where there is no conflict, the various embodiments of this disclosure and the features thereof in the embodiments may be combined with each other.

[0028] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to limit the disclosure. The term "and / or" as used in this disclosure includes any and all combinations of one or more of the associated enumerated entries. The singular forms "a" and "the" as used in this disclosure are also intended to include the plural forms, unless the context clearly indicates otherwise. The terms "comprising," "made of," etc., as used in this disclosure specify the presence of features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof.

[0029] Unless otherwise specified, all terms used in this disclosure (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this disclosure, and will not be interpreted as having an idealized or overly formal meaning, unless expressly so defined in this disclosure.

[0030] Communication failure data has the following key characteristics:

[0031] In communication networks, device failures exhibit complex interrelationships. A failure in one device not only affects the device itself but can also impact other devices through interconnections. The strength and health status of connections between devices are key factors influencing fault propagation. Therefore, effectively modeling the interconnections and fault states between devices is a significant challenge in assessing fault propagation.

[0032] Equipment failures often exhibit complex causal chains. A failure in one piece of equipment can lead to failures in other equipment, creating a chain reaction. The causal relationships between equipment are not limited to direct impacts but can also propagate indirectly through the connections between them. Identifying these causal chains is crucial for root cause analysis and future failure prediction. However, existing failure data analysis methods struggle to accurately infer complex causal relationships, resulting in incomplete explanations of failure modes.

[0033] Based on the characteristics of communication failure data, the industry faces the following pain points when extracting triples from knowledge graphs:

[0034] 1) Lack of effective data modeling mechanisms: Most triple extraction methods for related fault data analysis rely on an explicit perspective and fail to effectively integrate the connection relationships between devices, fault states, etc. For the extraction of complex fault propagation and causal chains, the relevant methods lack flexible data modeling mechanisms and are unable to effectively cope with the complex changes in relationships between devices.

[0035] 2) Difficulty in mining non-explicit knowledge: Communication fault data contains a large amount of non-explicit information. For example, coherent information such as causal relationships between devices and potential propagation paths of faults (coherence refers to the correlation between two or more physical quantities in time and space) often lacks clear labels or explicit features. Related triplet extraction methods usually struggle to effectively infer meaningful knowledge from this explicit information, resulting in extracted knowledge lacking depth and practical guiding significance.

[0036] To address the above issues, this embodiment obtains a network-corresponding device association model. This model represents the connection strength between multiple communication devices in the network and the device status of each device. Based on the device association model, multi-dimensional reasoning is performed on the communication devices to obtain multi-dimensional reasoning results. Based on the device association model and the multi-dimensional reasoning results, relationships are extracted from multiple communication devices to obtain a device knowledge graph in the network. This embodiment, by performing multi-dimensional reasoning on communication devices based on the device association model, can accurately capture the complex state relationships between devices in the communication network. It overcomes the shortcomings of traditional methods in dealing with state propagation and causal chain analysis, effectively handles complex relationship changes between devices, extracts data with depth and practical guiding significance, provides a clear and compact data foundation for subsequent reasoning, and ensures the accurate expression and efficient storage of event data.

[0037] The solutions disclosed herein can be applied to, but are not limited to, fields such as communication network management, device management, and fault diagnosis. They can be used in various application scenarios that require data extraction, such as, but not limited to, fault analysis, intelligent question answering, and fault prediction. The solutions disclosed herein can be used as a tuple (e.g., triple) extraction solution before creating a knowledge graph.

[0038] The embodiments of this disclosure will be described in detail below.

[0039] This disclosure provides a method for acquiring device data, such as... Figure 1 , Figure 2 As shown, it includes steps S11-S13:

[0040] S11. Obtain the network-corresponding device association model.

[0041] In this embodiment of the disclosure, the network is a mesh structure that connects devices such as PCs (Personal Computers), servers, and mobile phones through transmission media such as network cables and optical fibers, and transmission devices such as routers, switches, firewalls, and wireless access points (APs). This network may include, but is not limited to, a communication network. The communication network may include, but is not limited to, communication devices such as servers, routers, and switches.

[0042] In this embodiment of the disclosure, the device association model refers to a mathematical model that reflects the association relationship between multiple communication devices in a network.

[0043] In this embodiment of the disclosure, the device association model can be used to represent the connection strength between multiple communication devices in a network and the device status of each communication device.

[0044] In this embodiment of the disclosure, the device association model may be represented in the form of, but is not limited to, matrices, tables, etc.

[0045] In the embodiments disclosed herein, such as Figure 3 As shown, obtaining the network-corresponding device association model may include steps S21-S23:

[0046] S21. Obtain the physical topology characteristic data of communication devices in the network.

[0047] In this embodiment of the disclosure, physical topology can refer to the layout of network devices, the interconnection between devices, the geometry formed by the connected devices, etc. Physical topology feature data can refer to state data related to the physical topology of a device generated when a certain state occurs on the network device.

[0048] In this embodiment of the disclosure, a network monitoring system can be used to obtain status data (e.g., fault data) of communication devices and extract physical topology feature data from the status data.

[0049] In this embodiment of the disclosure, the physical topology feature data may include, but is not limited to: device identifier, device status, and connection strength.

[0050] S22. Extract the device identifier, device status, and connection strength between multiple communication devices from the physical topology feature data. In this embodiment of the disclosure, the device identifier, device status, and connection strength between multiple communication devices can be extracted from the physical topology feature data.

[0051] In this embodiment of the disclosure, the device identifier refers to the unique identifier of each communication device in the network, which can be used as an index of the device in the network. For example, the device identifier can be 1, 2, 3, ... k, where k is a natural number.

[0052] In this embodiment of the disclosure, the device state can be any state of the communication device. For example, the device state can include, but is not limited to, the fault state or health state of the communication device.

[0053] In this embodiment of the disclosure, the connection relationship between devices may include, but is not limited to, the connection strength between communication devices.

[0054] In this embodiment of the disclosure, the acquired physical topology feature data can first undergo data preprocessing. This data preprocessing may include, but is not limited to, data cleaning and data standardization.

[0055] In this embodiment of the disclosure, data cleaning may include, but is not limited to: removing noisy data and invalid data and filling in missing values ​​to ensure the integrity and consistency of the data.

[0056] In this embodiment of the disclosure, data standardization may include, but is not limited to: standardizing information such as device status and connection strength; for example, normalizing connection strength and device status values ​​to a unified range (such as [0,1]) so as to facilitate unified data formatting in subsequent steps; converting the physical location and network topology information of the device into a connection strength matrix, and converting the device fault status into a standardized category, etc.

[0057] S23. Establish a device association model based on device identification, device status, and connection strength.

[0058] In this embodiment of the disclosure, the device association model may include, but is not limited to, a multidimensional data matrix and a data table.

[0059] In this embodiment of the disclosure, the device association model can be a three-dimensional data matrix.

[0060] In this embodiment of the disclosure, when the device association model is a three-dimensional data matrix, establishing the device association model based on physical topology feature data may include:

[0061] Set the first dimension of the three-dimensional data matrix as the device identifier;

[0062] Set the second dimension in the three-dimensional data matrix as the connection strength;

[0063] Set the third dimension of the three-dimensional data matrix to the device state.

[0064] In this embodiment of the disclosure, the physical topology feature data of the hardware and software communication devices are extracted into a multidimensional data matrix M using a preset algorithm (e.g., including but not limited to the Apriori algorithm and its improved algorithms, rule-based and heuristic extraction algorithms, etc.) or by using a large model. xyz M xyz M is an n×n×m matrix (where n and m are both positive integers). xyz This represents the connection strength and device state (e.g., fault state) of communication devices x and y under a certain device state and its corresponding state category z, where x, y ∈ [1, n]: the index of the communication device (the number of communication devices is n), and z ∈ [1, m]: the index of the state (the number of states is m, e.g., the number of fault states is m). Where:

[0065] The first dimension (communication device dimension) represents the identifiers between communication devices, that is, the index of the communication device in the network.

[0066] The second dimension (connection strength dimension): represents the connection strength or distance between communication devices. Connection Strength D xy D represents the connection strength between communication device x and communication device y. xy A larger value indicates a closer relationship, a stronger connection, and a higher probability of state propagation (such as fault propagation).

[0067] The nth dimension (state dimension) represents the specific state category z of the communication device under different device states (such as fault states). Each communication device can have multiple state categories; for example, the fault state category could include normal, fault, and pending repair. The nth dimension of the matrix records the specific details of the communication device under different states.

[0068] In this embodiment of the disclosure, the multidimensional data matrix structure effectively integrates multidimensional information such as device identification, device status, and connection strength, providing a clear data foundation for the construction of knowledge graphs.

[0069] In this embodiment of the disclosure, the constructed multidimensional data matrix can be dynamically updated.

[0070] In this embodiment of the disclosure, dynamic updates may include, but are not limited to, dynamically updating the data in the device association model by monitoring changes in device status and connection strength in real time. When the device status or connection strength changes, the corresponding values ​​in the device association model are updated in real time to ensure that the device association model can reflect the latest state of the network.

[0071] In this embodiment of the disclosure, by modeling the physical topology feature data (e.g., device identifier, connection strength, and device status) of hardware and software communication devices in the form of a device association model, a structured multidimensional relationship data (device identifier, connection strength, and device status) between communication network topology devices is provided for subsequent root cause reasoning and propagation reasoning of device status (e.g., fault status). This provides more contextual information for state coherence reasoning and reduces computational complexity based on efficient access to multidimensional relationship data.

[0072] S12. Perform multi-dimensional reasoning on communication devices based on the device association model to obtain multi-dimensional reasoning results.

[0073] Causal reasoning is a crucial technique in modern data analysis and artificial intelligence, widely applied to identifying causal relationships and modeling causal propagation processes in complex systems. In fields such as communication networks, equipment management, and fault diagnosis, the propagation of equipment faults is often not a simple direct impact, but rather affects other equipment through complex connections and causal chains between devices. Therefore, traditional state analysis methods based on simple rules cannot accurately reveal the deep-seated causal relationships and state propagation paths between devices.

[0074] In this embodiment of the disclosure, the multidimensional reasoning result may include, but is not limited to, causal reasoning result and / or propagational reasoning result.

[0075] In this embodiment of the disclosure, multi-dimensional reasoning of communication devices based on a device association model to obtain multi-dimensional reasoning results may include:

[0076] Based on a device association model, the strength of causal relationships between multiple communication devices is dynamically inferred, and the causal inference results are output; and / or,

[0077] Based on the device association model, the strength of causal relationships between multiple communication devices is dynamically inferred, and the causal inference results are output. Based on the multidimensional data matrix and the causal inference results, the probability of device state propagating from one communication device to another is calculated, and the propagation inference results are output.

[0078] In this embodiment, the multidimensional reasoning is based on a device association model of the physical topology characteristics of hardware and software communication devices to perform causal reasoning and / or propagation reasoning. By analyzing the state similarity and connection strength between communication device states, the causal chain of event occurrence is inferred, and the strength of the causal relationship between communication devices (i.e., the causal reasoning result) is obtained. Furthermore, the propagation path of a state occurrence from one communication device to another is predicted, along with the transmission probability of that path (i.e., the propagation reasoning result). The combination of causal reasoning and propagation reasoning can deeply explore the relationship between state occurrence and propagation between direct and indirect nodes of communication devices, providing a more comprehensive and accurate fault analysis solution for state prediction (such as fault prediction), event risk management (such as fault risk management), and network event optimization. It is an effective solution for intelligent fault operation and maintenance and adaptive network fault management.

[0079] In this embodiment of the disclosure, state coherence information between communication devices can be obtained during the multidimensional reasoning process. This state coherence information may include, but is not limited to, state changes between communication devices (e.g., fault state changes), connection strength, and state similarity (e.g., fault state similarity).

[0080] In the embodiments disclosed herein, such as Figure 4 As shown, the strength of causal relationships between multiple communication devices is dynamically inferred based on the device association model, and the causal inference results are output, including steps S31-S33:

[0081] S31. Determine the state similarity between multiple communication devices based on the device states of multiple communication devices.

[0082] In this embodiment of the disclosure, state similarity is used to measure the similarity of multiple communication devices in terms of state, reflecting the degree of similarity of the state patterns (e.g., fault modes) of the communication devices, thereby facilitating subsequent steps to infer complex causal relationships between communication devices based on the state similarity.

[0083] S32. Adjust the connection strength according to the preset adjustment method to obtain the adjustment result.

[0084] In this embodiment of the disclosure, the adjustment result includes a first adjustment result and a second adjustment result; adjusting the connection strength according to a preset adjustment method to obtain the adjustment result may include:

[0085] A linear adjustment of the connection strength yields the first adjustment result; and,

[0086] The connection strength is dynamically adjusted based on preset dynamic adjustment parameters to obtain a second adjustment result.

[0087] In this embodiment of the disclosure, for example, linear adjustment can be performed using a preset linear function. D xy +β1 is a linear function used to linearly adjust the connection strength. This is to ensure that weaker connections have a smaller impact on the failure of communication device y in causal reasoning, while stronger connections amplify this impact. Linear adjustment function. Capable of measuring the connection strength D between communication devices xy Direct scaling is performed to control the influence of connection strength in causal inference. This scaling process can be flexibly controlled by adjusting γ1 and β1. By introducing a bias term β1, a baseline value can be set for the causal influence of all connections, preventing the complete suppression of causal influence when some connection strengths are zero. For the hyperparameters β1 and γ1, the optimal solution is obtained by constructing grid parameters using historical connection strengths between communication devices and the probability distribution data of communication device states as input features.

[0088] In this embodiment of the disclosure, a linear adjustment function is used. Further flexible adjustments to the connection strength ensure that the causal relationship of state propagation is enhanced only when the connection strength is strong, while weaker connections are appropriately suppressed.

[0089] In this embodiment of the disclosure, the dynamically adjusted parameter α1 represents the connection strength D. xy The degree of influence on causal relationships. The introduction of dynamically adjusted parameters allows the impact of connection strength on inter-fault coherence inference to weaken as connection strength decreases. Generally, the larger α1 is, the more significant the attenuation, meaning that more distant communication devices (i.e., weakly connected communication devices) contribute less to causal relationships, increasing sensitivity to nearby communication devices. For the α1 hyperparameter, historical connection strength between communication devices, historical device states, and the probability distribution data of communication device states (such as fault states) are used as input features, with random hyperparameter values ​​as a benchmark. The hyperparameter is automatically adjusted by minimizing the error between the predicted and actual values ​​of the input feature data. Its core is to optimize α1 through historical fault data, enabling the causal inference algorithm to adapt to changes in different network topologies and communication device states. With the goal of minimizing the prediction error, a loss function is defined, and gradient descent is used to iteratively update the causal inference algorithm until the loss function converges to the set target.

[0090] S33. Obtain the causal relationship strength based on the state similarity and adjustment results, and output the causal relationship strength as the causal reasoning result.

[0091] In this embodiment of the disclosure, the above-described causal reasoning process can be implemented using a preset causal reasoning algorithm, which may include:

[0092]

[0093] Among them, C xy S represents the strength of the causal relationship between the state of communication device x and the state of communication device y; x (z) and S y (z) represent the specific state categories of communication devices x and y, respectively; Sim(S x (z),S y (z) represents the state similarity between communication device x and communication device y; D xy This indicates the connection strength between communication device x and communication device y; This refers to linear adjustment of the connection strength; α1 is a dynamic adjustment parameter.

[0094] In this embodiment of the disclosure, by combining the state similarity, connection strength and linear adjustment mechanism between communication devices, the causal relationship between communication devices can be accurately derived.

[0095] In this embodiment of the disclosure, C xy C represents the strength of the causal relationship between the state of communication device x (e.g., a fault state) and the state of communication device y (e.g., a fault state). xy The larger the value, the higher the probability that the communication device y will experience this event (such as a failure).

[0096] In this embodiment of the disclosure, S x (z) and S y (z) represent the states (e.g., fault states) of communication devices x and y, respectively, which can be quantified by state similarity.

[0097] In this embodiment of the disclosure, the state similarity Sim(S) x (z),S y (z) measures the similarity between communication devices x and y in terms of their states, reflecting the degree of similarity in their state patterns (e.g., failure modes), thereby inferring complex causal relationships between the devices. For example, if both communication devices are in the same state category (e.g., both in a failure state, or both in a normal state, or both in a state awaiting repair), the similarity is 1; if the state categories are different, the similarity is 0.

[0098] In this embodiment of the disclosure, D xy (Connection strength) represents the dependency between communication device x and communication device y. Connection strength is typically determined by factors such as the physical distance between the communication devices and the network topology. xyThe larger the value, the closer the relationship between communication devices, that is, the stronger the connection strength, and the higher the possibility of state propagation between communication devices. The dynamic adjustment parameter α1 adjusts the influence of connection strength on causality, avoiding overemphasizing the influence of weak connections.

[0099] In this embodiment of the disclosure, through this comprehensive analysis of multiple factors and dimensions, it is possible to dynamically adapt to changes in the relationships between communication devices, more accurately identify causal chains and dynamic propagation paths between communication devices, and thus achieve more accurate state prediction (such as fault prediction) and event propagation analysis (such as fault propagation analysis) in complex communication networks.

[0100] In the embodiments of this disclosure, during the state coherence inference process between communication devices, a dynamic adjustment parameter α1 based on convergence calculation of historical data is introduced in the linear correction of connections between distant and weak communication devices. This achieves precise adjustment of the influence of connection strength on causal relationships, reduces the excessive influence of weak connections between communication devices, and ensures that the causal relationship is not excessively enhanced when there are connections between distant or weak communication devices. This allows the causal inference algorithm to accurately reflect the influence of strong connections while adapting to the influence of weak connections, and avoids them causing excessive interference to the causal inference results.

[0101] In this embodiment of the disclosure, the causal reasoning algorithm is based on:

[0102] Causal effect decay: The propagation of states (such as faults) follows the physical law that "the farther the distance, the smaller the impact," through... Achieve attenuation modeling;

[0103] State similarity hypothesis: If two communication devices are in highly similar states, their states (such as failures) may be caused by common factors or there may be a one-way causal chain.

[0104] Dynamic parameter optimization: Based on historical fault data, α1 and γ1 are optimized through grid search or gradient descent to minimize prediction error and ensure the generalization ability of the causal inference algorithm.

[0105] In this embodiment of the disclosure, the causal reasoning algorithm identifies deep causal chains and mines potential causal relationships between devices by using the state similarity and connection strength between communication devices, thereby providing accurate data for causal triples in the knowledge graph.

[0106] In the embodiments disclosed herein, such as Figure 5 As shown, the probability of a device state propagating from one communication device to another is determined based on the device association model and causal inference results, and the propagation inference results are output. This may include steps S41-S42:

[0107] S41. Determine the state similarity between multiple communication devices based on the device states of multiple communication devices.

[0108] In this embodiment of the disclosure, state similarity is used to measure the similarity of multiple communication devices in terms of state, reflecting the degree of similarity of the state patterns (e.g., fault modes) of the communication devices, thereby facilitating subsequent steps to infer complex causal relationships between communication devices based on the state similarity.

[0109] S42. Obtain the probability based on state similarity, connection strength, and causal reasoning results, and output the probability as the propagation reasoning result.

[0110] In this embodiment of the disclosure, the above-described propagation inference process can be implemented by a preset propagation inference algorithm, which may include:

[0111] P xy =σ(α2·Sim(S) x (z),S y (z))+β2·D xy +γ2·C xy (2)

[0112] Among them, P xy σ represents the probability of communication device x propagating its state to communication device y; σ is the activation function; α² is the similarity weight; β² is the connection strength weight; and γ² is the causal influence weight.

[0113] In this embodiment of the disclosure, the propagation path of an event between communication devices can be predicted based on the state similarity, connection strength, and connection strength between communication devices.

[0114] In this embodiment of the disclosure, the state similarity Sim(S) x (z),S y (z) measures the propagation probability of communication devices x and y when the communication devices are in a certain device state. It represents the degree of similarity in the occurrence of states. The higher the state similarity, the greater the possibility of event propagation.

[0115] In this embodiment of the disclosure, D xy This reflects the strength of the physical or logical connection between communication devices; the stronger the connection, the higher the probability of state propagation. Causality strength C xy This indicates the potential impact of faults between communication devices, and the probability of event propagation is influenced by a combination of state similarity and connection strength.

[0116] In this embodiment of the disclosure, the propagation inference algorithm introduces hyperparameters α2, β2 and γ2 to adjust the contribution of state similarity, connection strength and causal relationship strength to the propagation probability, respectively, to ensure that each factor has a reasonable weight in the propagation inference.

[0117] In this embodiment of the disclosure, through a multi-dimensional reasoning approach, the propagation reasoning algorithm can accurately predict the probability of a device state propagating from one communication device to another, identify state (such as a fault) propagation chains, and provide a scientific basis for state (such as a fault) prevention and control.

[0118] In this embodiment of the disclosure, σ is used as an activation function to map the propagation probability to the range [0,1].

[0119] In this embodiment, for hyperparameters α2, β2, and γ2, historical connection strength, historical device states, and probability distribution data of device states occurring between communication devices (probability of state occurrence and propagation path information between communication devices) are used as input features. Random hyperparameter values ​​are used as a benchmark. The hyperparameters are automatically adjusted by minimizing the error between the predicted and actual values ​​of the input feature data. These are used to adjust the contribution of state similarity, connection strength, and causal influence to the propagation probability. α2 (State Similarity Weight): When the similarity of the fault states of communication devices is high, the probability of propagating the state is greater. If α2 is too large, it will overemphasize the similarity of device states, potentially leading to misjudgment. Conversely, if it is too small, it may ignore the state propagation relationship between similar communication devices. β2 (Connection Strength Weight): Connection strength is an important factor determining the dependency relationship between communication devices. Too large or too small a weight may lead to deviations in fault propagation prediction. Appropriately adjusting the value of β2 can better reflect the contribution of the physical connection strength between communication devices to the propagation probability. γ² (Causal Influence Weight): The causal relationship influences the path of state propagation between communication devices. A high γ² weight indicates a strong focus on the causal chain between communication devices. Appropriate causal influence weights can help identify potential state propagation paths between communication devices, avoiding overly simplistic predictions based solely on connection strength.

[0120] In this embodiment of the disclosure, the theoretical basis of the propagation inference algorithm may include:

[0121] Multi-factor fusion: Propagation probability is determined by direct connection (D xy ), State similarity (Sim(S) x ,S y )) and indirect causal chain (C xy This decision was made jointly, covering both explicit and implicit transmission pathways;

[0122] Nonlinear activation function: The σ function transforms the linear combination result into a probability, which conforms to the threshold characteristics of fault propagation (such as "critical connection strength" triggering propagation);

[0123] Dynamic adaptability: By updating the device status and connection strength in the multidimensional data matrix in real time, the propagation inference algorithm can dynamically adjust the propagation probability to adapt to changes in network topology.

[0124] In this embodiment of the disclosure, the propagation inference algorithm combines the connection strength, state similarity and causal relationship between communication devices to predict the path and probability of fault propagation, thereby further enriching the triples of the knowledge graph.

[0125] S13. Based on the device association model and multidimensional reasoning results, the relationships of multiple communication devices are extracted to obtain the device knowledge graph in the network.

[0126] In this embodiment of the disclosure, the multidimensional reasoning result may include: causal reasoning result and propagational reasoning result.

[0127] Based on the device association model and multidimensional reasoning results, relationships are extracted from multiple communication devices to obtain a device knowledge graph in the network, which may include:

[0128] Basic tuples are extracted based on the device association model; these basic tuples are used to represent the device status of multiple communication devices.

[0129] Causal tuples are extracted based on the results of causal inference; causal tuples are used to represent the probability of causal influence between the states of multiple communication devices.

[0130] Based on the propagation inference results, propagation tuples are extracted; propagation tuples are used to represent the probability that a device state propagates from one communication device to another.

[0131] A device knowledge graph is constructed based on basic tuples, causal tuples, and propagation tuples.

[0132] In this embodiment of the disclosure, based on the device association model described above, and the causal reasoning results and propagation reasoning results obtained through state coherence reasoning, the following tuple information of state information can be obtained:

[0133] Basic triples: Information about communication devices that are affected by an event in the network topology deployment. Basic triples (e.g., triples) are obtained through the device states and connection strengths between devices in the device association model. For example: (A, B, 2): This indicates that after fault S occurs, both communication devices A and B in the network topology are affected, and the connection strength between communication devices A and B is 2.

[0134] Causal triples: Based on the device states and connection strengths between communication devices in the device association model, causal reasoning is used to derive the antecedent-effect relationships between the device states of communication devices. For example: (A, B, 0.6): indicates that a failure of communication device A has a 60% probability of causing a failure of communication device B.

[0135] Propagation triples: Based on the connection strength between communication devices in the device association model and the propagation probability inference of causal tuple data, this yields the probability that a certain device state will propagate from one communication device to another after it occurs. For example: (A, B, 0.83): indicates that the probability of a fault in communication device A propagating to communication device B is 83%.

[0136] In this embodiment, dynamic adaptation to changes in the relationships between communication devices is achieved, enabling real-time updates of the device graph in complex network environments. This improves the accuracy of fault prediction and the timeliness of the knowledge graph, providing an efficient and accurate knowledge graph construction and reasoning method for intelligent diagnosis and prevention of communication network faults, and has significant application value.

[0137] In this embodiment of the disclosure, in order to more clearly illustrate the modeling of the device association model, the calculation of the causal reasoning algorithm, and the calculation of the propagation reasoning algorithm, a simplified example is used below to illustrate the embodiment of this application.

[0138] Suppose there is a simple communication network containing three communication devices (e.g., communication device A, communication device B, and communication device C), each of which can be in a different fault state category (e.g., normal, faulty, awaiting repair). The following demonstrates how to implement device association modeling, causal reasoning calculation, and propagation reasoning calculation.

[0139] In this disclosure, we first introduce an embodiment of device association modeling.

[0140] Assume the connection strengths between three communication devices (e.g., communication device A, communication device B, and communication device C) are as shown in Table 1 below (a higher connection strength indicates a closer relationship between the communication devices):

[0141] Table 1

[0142]

[0143]

[0144] The values ​​in Table 1 represent the connection strength between communication devices. For example, the connection strength between communication device A and communication device B is 2, and the connection strength between communication device A and communication device C is 0, indicating that there is no direct strong connection between communication device A and communication device C.

[0145] Assume the fault status indicators of the communication equipment are as follows: Normal = 0; Fault = 1; Need to be repaired = 2.

[0146] The current fault status of communication equipment A, communication equipment B, and communication equipment C is shown in Table 2:

[0147] Table 2

[0148] Communication equipment A B C Fault status 1 0 2

[0149] That is, communication device A is in a faulty state, communication device B is in a normal state, and communication device C is in a state awaiting repair.

[0150] Based on the data in Tables 1 and 2, a device association model, such as a multidimensional data matrix M, can be constructed to represent the relationships between communication devices. This multidimensional data matrix M can be a three-dimensional data matrix with three dimensions:

[0151] Communication device dimension: Represents communication devices A, B, and C;

[0152] Connection strength dimension: Represents the connection strength between communication devices;

[0153] State dimension: Represents the state z of each communication device under different fault states.

[0154] The multidimensional data matrix M can be simplified as follows:

[0155] Connection strength of communication equipment:

[0156]

[0157] Fault status of communication equipment:

[0158] M stat =[1 0 2]; (4)

[0159] The specific representation of a multidimensional matrix:

[0160]

[0161] In this disclosure, embodiments of the causal reasoning scheme are described below.

[0162] Assume that the fault status of communication device A is "fault" (1), and the fault status of communication device B is "normal" (0).

[0163] The connection strength information between communication devices is as follows:

[0164] Connection strength D AB =2 (This represents the connection strength between communication device A and communication device B; the smaller the value, the stronger the connection).

[0165] The steps for calculating the strength of causality may include steps 1-4:

[0166] Step 1: Calculate the similarity Sim(S) between the fault states of communication device A and communication device B. A (z),SB (z)).

[0167] Fault state S of communication device A A (z) = 1, Fault state S of communication device B B (z) = 0.

[0168] Hamming distance can be used to calculate the similarity of fault states between communication device A and communication device B. The similarity calculation for their fault states is as follows:

[0169]

[0170] so:

[0171]

[0172] Step 2: Calculate the connection strength D between communication devices. AB .

[0173] The connection strength between communication device A and communication device B, as shown in the multidimensional data matrix, is D. AB =2.

[0174] Step 3: Calculate the function

[0175] By definition, a function The strength used to adjust causality based on the connection strength between communication devices. Assuming γ1 = 0.25 and β1 = 0:

[0176]

[0177] Step 4: Calculate the causal relationship strength Cxy.

[0178] According to the calculation formula:

[0179]

[0180] Assuming the dynamic adjustment parameter α1 is 1, substitute the value calculated in the above steps:

[0181]

[0182] Therefore, the causal relationship strength (or causal influence strength) of communication device A's failure on communication device B is 0.125 (i.e., the causal reasoning result).

[0183] In this embodiment of the disclosure, based on the causal reasoning result, the causal triple can be extracted as (A, B, 0.125).

[0184] In this embodiment of the disclosure, the propagation path reasoning scheme can be implemented based on the data in the multidimensional data matrix and the propagation reasoning algorithm, which will not be elaborated here. Based on the obtained propagation reasoning results, propagation triples can be extracted.

[0185] In this embodiment of the disclosure, the tuple data obtained based on the scheme of this embodiment of the disclosure can be used to construct an event knowledge graph.

[0186] In this disclosure, an embodiment of constructing an event knowledge graph (e.g., a fault knowledge graph) based on multi-group data obtained from the scheme of this application is given below.

[0187] In this embodiment of the disclosure, the core elements of the fault knowledge graph can first be defined:

[0188] Entity (Node): Represents a communication device in a communication network (such as communication device A, communication device B). Attributes include device identifier (ID), current device status (normal / faulty / pending repair), historical fault records, etc.

[0189] Relationship (edge):

[0190] - Basic Relationships: Describes the physical or logical connection between communication devices; attributes include connection strength (D). xy ).

[0191] - Causality: Indicates the strength of the impact of faults between communication devices; the attribute is the causality strength value (C). xy ).

[0192] - Propagation relationship: Represents the fault propagation path and probability; the attribute is the propagation probability (P). xy ).

[0193] In this embodiment of the disclosure, the graph pattern of the fault knowledge graph can then be determined:

[0194] Node type: communication equipment, fault type, network topology area.

[0195] Edge types: connection relationship, causal relationship, propagation path.

[0196] Attribute example:

[0197] - Device node attributes: {ID:"x", Status:"Fault", Historical fault count:3,...};

[0198] - Connection relationship attributes: {Connection strength: 2, Update time: "2023-10-01",...};

[0199] - Causal relationship attributes: {Causal strength: 0.6, Confidence: 0.85, ...};

[0200] - Propagation path attributes: {Propagation probability: 0.83, Propagation direction: "x→y", ...}.

[0201] In this embodiment of the disclosure, a fault knowledge graph is finally created based on the above core elements and graph patterns:

[0202] 1) Data source input

[0203] The input data includes: device association model and tuple data.

[0204] Device association model: includes device identifier, connection strength (D) xy ), Equipment Status (S) x (z)).

[0205] Multivariate data: Basic triples (x, y, D) xy ), causal triples (x, y, C) xy ), propagation triples (x, y, P) xy ), or can be represented as: basic triples (x, D) xy , y), causal triples (x, C) xy , y), propagation triples (x, P) xy ,y).

[0206] 2) The attribute of triples mapped to edges

[0207] For the basic triple: (communication device x, connection relationship, communication device y) → attribute: {connection strength: 2};

[0208] For the causal triple: (communication device x, causal relationship, communication device y) → attribute: {causal relationship strength: 0.6};

[0209] For the propagation triple: (communication device x, propagation path, communication device y) → attribute: {propagation probability: 0.83}.

[0210] 3) Graph database storage and indexing

[0211] Storage engine: Uses graph databases (such as Neo4j and NebulaGraph) to store triple data, supporting efficient relation queries and path analysis.

[0212] Indexing strategy:

[0213] Establish node indexes for communication device IDs and fault status to accelerate device status queries.

[0214] Build edge attribute indexes for connection strength, causal relationship strength, and propagation probability, and support filtering by threshold (e.g., querying paths with propagation probability > 0.5).

[0215] 4) Dynamic update mechanism

[0216] - Real-time data synchronization: When the device status or connection strength in the device association model changes, the following process is triggered:

[0217] a. Update the corresponding data in the device association model;

[0218] b. Re-execute the causal reasoning algorithm and the propagation reasoning algorithm to generate new triples;

[0219] c. Incrementally update the node and edge attributes in the graph database.

[0220] - Historical data archiving: Regularly archive old data to a time-series database (such as InfluxDB) to ensure the lightweight nature and query efficiency of the knowledge graph.

[0221] In this embodiment of the disclosure, the application scenarios of fault knowledge graphs may include, but are not limited to:

[0222] 1) Fault cause location

[0223] Query example: Search the fault knowledge graph for all causal relationship edges associated with the faulty device, sort them in descending order of causal strength, and quickly locate the most likely source of the fault.

[0224] MATCH(a:device)-[r:causal relationship]->(b:device{state:"fault"});

[0225] RETURN a.ID,r.causal intensity ORDER BY r.causal intensity DESC.

[0226] 2) Fault propagation path prediction

[0227] Path analysis: Identifying high-risk transmission chains based on propagation probability.

[0228] MATCH path = (a: device) - [: propagation path * 1..3] -> (b: device)

[0229] WHERE ALL(r IN relationships(path)WHERE r. propagation probability>0.7)

[0230] RETURN path

[0231] 3) Intelligent question answering and decision support

[0232] Natural Language Query: Through the API (Application Programming Interface) of the graph database, the natural language questions from operations and maintenance personnel (such as "Which devices may be affected by the failure of device A?") are converted into fault knowledge graph query statements and returned in a structured manner.

[0233] The present disclosure includes at least the following advantages:

[0234] 1. Traditional fault reasoning methods typically rely on simple explicit rules or analysis based on two-dimensional or 0-1 modeling, lacking in-depth exploration of complex causal relationships between devices. The present invention introduces a linear adjustment mechanism for device state similarity and connection strength to deduce complex causal chains between devices, significantly improving the efficiency of fault triple extraction from communication network devices. This is particularly effective when dealing with complex relationships between communication devices and fault propagation paths, enhancing the ability to analyze fault causes and predict fault propagation.

[0235] 2. The solution in this disclosure uses a device association model to accurately identify causal relationships between communication devices. Unlike traditional rule-based or simple modeling-based reasoning methods, the algorithm in this disclosure not only comprehensively considers device identification, device status, and device connection strength, but also introduces dynamic adjustment parameters and a linear adjustment mechanism for connection strength. Based on changes in device status, it captures the probability of fault occurrence and the antecedents and consequences of propagation paths in hardware and software devices. In scenarios with multi-dimensional and complex topologies of hardware and software device deployments, fault propagation paths are long and have multiple impacts. Coupled with dynamic changes in topology health, the combination of dynamic adjustment parameters and linear functions to adjust the weight of connection strength ensures that fault causal reasoning can flexibly respond to changes in network topology and device status, more accurately inferring the propagation and impact of faults in indirectly connected hardware and software devices with multiple nodes at intervals. This deep causal chain mining significantly improves the accuracy and scope of fault analysis and can identify indirect associations and complex fault modes that are ignored by traditional methods.

[0236] 3. The solution of this disclosure also accurately predicts the propagation path and probability of faults between devices by analyzing the connection strength, fault state similarity, and causal relationship strength between devices. Traditional methods often only focus on the local impact between devices, ignoring the complex propagation chain between devices, while the solution of this disclosure can accurately identify the probability of a fault propagating from one device to another. Especially when the connection strength and state similarity between devices are unbalanced, a weighting mechanism is introduced to flexibly adjust the propagation path prediction based on the connection strength, causal influence, and state similarity between devices, thereby providing a scientific basis for fault prevention and maintenance decisions.

[0237] 4. The embodiments of this disclosure significantly improve the efficiency of knowledge graph construction for communication network equipment faults by using device association modeling and innovative causal and propagation reasoning algorithms. The multi-dimensional matrix structure effectively integrates multi-dimensional information such as device fault status, connection strength, and health status, providing a clear data foundation for knowledge graph construction. The causal reasoning algorithm identifies deep-seated causal chains and mines potential causal relationships between devices by calculating the state similarity and connection strength between devices, thus providing accurate data for causal triples in the knowledge graph. The propagation reasoning algorithm combines the connection strength, state similarity, and causal relationships between devices to predict the path and probability of fault propagation, further enriching the triples in the knowledge graph. This method dynamically adapts to changes in inter-device relationships, enabling real-time updates of the knowledge graph in complex network environments, improving the accuracy of fault prediction and the timeliness of the knowledge graph. In summary, the embodiments of this disclosure provide an efficient and accurate knowledge graph construction and reasoning method for intelligent diagnosis and prevention of communication network faults, possessing significant application value.

[0238] 5. The solution in this disclosure combines causal prediction and propagation prediction to deeply mine complex fault modes and propagation paths among devices, constructing a precise equipment fault knowledge graph. This allows maintenance personnel to quickly locate fault sources, identify potential fault chains, and provide data support for subsequent fault prevention, equipment scheduling, and resource allocation. Furthermore, leveraging causal and propagation reasoning capabilities, the system can issue early warnings before faults occur, enabling early fault prediction and rapid response. The intelligent query function of the knowledge graph provides root cause analysis and dynamic emergency response suggestions, improving network reliability and stability.

[0239] In the embodiments disclosed herein, significant advantages are demonstrated in fault analysis, intelligent question answering, and fault prediction application scenarios:

[0240] Fault Analysis: By using device association modeling and causal reasoning algorithms, we can deeply explore the complex causal relationships between devices in the communication network and accurately identify the root cause of faults. Combining the connection strength and state similarity between devices, we can provide a clear causal chain for fault analysis, helping maintenance personnel quickly locate the source of problems, improve fault diagnosis efficiency, and contribute to AI fault assistant analysis.

[0241] Intelligent Question Answering: Knowledge graph construction capabilities enable the presentation of equipment fault-related information in a structured format. Through the intelligent question answering system, maintenance personnel can quickly query information such as the fault status, causal relationships, and fault propagation paths between devices, providing real-time and accurate fault solutions and improving decision-making efficiency.

[0242] Fault Prediction: Utilizing propagation inference algorithms, the propagation path and probability of equipment faults can be accurately predicted, providing a scientific basis for fault early warning and prevention. By dynamically adjusting the inference model, it can adapt to changes in equipment relationships in real time, effectively reducing the impact of equipment faults on the overall network, improving system stability and reliability, and enabling applications in fault-related digital twin systems.

[0243] This disclosure also provides an electronic device 100, such as... Figure 6 As shown, it includes:

[0244] One or more processors 101;

[0245] The memory 102 stores one or more programs, which, when executed by one or more processors, enable one or more processors 101 to implement the device data acquisition method.

[0246] One or more input / output I / O interfaces 103 are connected between the processor 101 and the memory 102 and configured to enable information exchange between the processor 101 and the memory 102.

[0247] This disclosure also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the device data acquisition method.

[0248] This disclosure also provides a computer program product, which includes a computer program that, when executed by a processor, implements the device data acquisition method.

[0249] In the embodiments disclosed herein, any of the foregoing embodiments of the memory management method are applicable to the embodiments of the electronic device, storage medium, and program product, and will not be described in detail here.

[0250] Those skilled in the art will understand that all or some of the functional modules / units disclosed above can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0251] In hardware implementations, the division between functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be executed by several physical components working together.

[0252] Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit (CPU), digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technique for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory (FLASH) or other disk storage; read-only optical disc (CD-ROM), digital versatile disc (DVD) or other optical disc storage; magnetic cartridges, magnetic tapes, disk storage or other magnetic storage; and any other media that can be used to store desired information and can be accessed by a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0253] This disclosure has disclosed exemplary embodiments, and although specific terminology has been used, it is for general illustrative purposes only and should not be construed as limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in conjunction with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in conjunction with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of this disclosure as set forth by the appended claims.

Claims

1. A method for acquiring device data, comprising: Obtain the network-corresponding device association model, which is used to represent the connection strength between multiple communication devices in the network and the device status of each communication device; Based on the device association model, multidimensional reasoning is performed on the communication device to obtain multidimensional reasoning results; Based on the device association model and the multidimensional reasoning results, the relationships of the multiple communication devices are extracted to obtain the device knowledge graph in the network.

2. The device data acquisition method according to claim 1, wherein, The multidimensional reasoning of the communication device based on the device association model to obtain the multidimensional reasoning result includes: Based on the device association model, the strength of causal relationships among multiple communication devices is dynamically inferred, and the causal inference results are output; and / or, Based on the device association model, the strength of causal relationships between multiple communication devices is dynamically inferred, and the causal inference result is output. Based on the device association model and the causal inference result, the probability of the device state propagating from one communication device to another is determined, and the propagation inference result is output.

3. The device data acquisition method according to claim 1, wherein, The process of obtaining the network-corresponding device association model includes: Obtain the physical topology characteristic data of the communication devices in the network; Extract the device identifier, device status, and connection strength between the multiple communication devices from the physical topology feature data; The device association model is established based on the device identifier, the device status, and the connection strength.

4. The device data acquisition method according to claim 1, wherein, The multidimensional reasoning results include: causal reasoning results and propagational reasoning results; The step of extracting relationships between the multiple communication devices based on the device association model and the multidimensional reasoning results to obtain a device knowledge graph in the network includes: Based on the device association model, basic tuples are extracted; the basic tuples are used to represent the device states of multiple communication devices. Based on the causal inference results, causal tuples are extracted; the causal tuples are used to represent the causal influence probability of the device states among multiple communication devices; Based on the propagation inference results, a propagation tuple is extracted; the propagation tuple is used to represent the probability that the device state propagates from one communication device to another. The device knowledge graph is constructed based on the basic tuples, the causal tuples, and the propagation tuples.

5. The device data acquisition method according to claim 2, wherein, The dynamic inference of the causal relationship strength between multiple communication devices based on the device association model, and the output of the causal inference result, includes: Determine the state similarity among the multiple communication devices based on their device states; The connection strength is adjusted according to a preset adjustment method to obtain the adjustment result; The causal relationship strength is obtained based on the state similarity and the adjustment result, and the causal relationship strength is output as the causal reasoning result.

6. The device data acquisition method according to claim 5, wherein, The adjustment results include a first adjustment result and a second adjustment result; The step of adjusting the connection strength according to a preset adjustment method to obtain the adjustment result includes: The connection strength is linearly adjusted to obtain the first adjustment result; as well as, The connection strength is dynamically adjusted based on preset dynamic adjustment parameters to obtain the second adjustment result.

7. The device data acquisition method according to claim 2, wherein, The process of determining the probability of a device state propagating from one communication device to another based on the device association model and the causal inference result, and outputting the propagation inference result, includes: Determine the state similarity among the multiple communication devices based on their device states; The probability is obtained based on the state similarity, the connection strength, and the causal reasoning result, and the probability is output as the propagation reasoning result.

8. The device data acquisition method according to claim 3, wherein, The device association model is a three-dimensional data matrix; The process of establishing the device association model based on the device identifier, the device status, and the connection strength includes: Set the first dimension of the three-dimensional data matrix as the device identifier; Set the second dimension of the three-dimensional data matrix as the connection strength; Set the third dimension of the three-dimensional data matrix to the device state.

9. An electronic device, comprising: One or more processors; A memory having stored one or more programs thereon, which, when executed by one or more processors, cause the one or more processors to implement the device data acquisition method according to any one of claims 1-8; One or more input / output (I / O) interfaces are connected between the processor and the memory and configured to enable information exchange between the processor and the memory.

10. A computer program product comprising a computer program that, when executed by a processor, implements the device data acquisition method according to any one of claims 1-8.

Images