Communication room monitoring method and system based on big data

Abstract

The present application provides a communication room monitoring method and system based on big data, including: obtaining environmental time series data, equipment operation time series data and inspection frequency time series data of the communication room in a historical time period; obtaining estimated environmental data, estimated equipment operation data and estimated inspection frequency data of the communication room in a future set time period; fusing the estimated environmental data, estimated equipment operation data and estimated inspection frequency data, and constructing a security data set; evaluating the data priority of the security data set, obtaining a priority evaluation result, and establishing a subcontracting data set; performing fault prediction; thereby enhancing the stability and interpretability of the monitoring results, achieving accurate fault prediction for the communication room, improving fault prediction efficiency, and facilitating the safe management and stable operation of the communication room.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a communication room monitoring method and system based on big data. Background Art

[0002] With the rapid development of social informatization, communication rooms are increasingly used in various industries, becoming a key link in ensuring the quality of communication services and maintaining the normal operation of communication equipment and systems. However, communication rooms are also susceptible to external environmental factors and equipment abnormalities. During operation, equipment failures such as equipment heating, excessive temperature and humidity, and abnormal alarms are prone to occur, which greatly affects the reliability and availability of communication rooms. Therefore, real-time monitoring of equipment operation and room environment status to predict equipment failures has become an important method to ensure the reliability of communication rooms. Collecting historical data, mining key influencing factors, and predicting faults in communication rooms are of great significance in practical applications in the field of communications.

[0003] In the existing technology, firstly, the traditional methods ignore the problem of limiting the accuracy of data prediction, and often fail to distinguish between correlation and true causality, making the final monitoring results unstable and lacking in interpretability; secondly, the existing methods do not recognize that certain factors may have a disproportionately greater impact on the safety of the computer room. Since the key information in the data is not fully utilized during the monitoring process, it is difficult for traditional methods to achieve accurate fault prediction; thirdly, the importance of different data is different, and traditional methods usually lack analysis of related data, resulting in untimely processing of important data and inefficient fault prediction.

[0004] Therefore, developing a communication room monitoring method and system based on big data has important practical significance and broad application prospects. Summary of the invention

[0005] The purpose of this application is to provide a communication room monitoring method and system based on big data, so as to enhance the stability and interpretability of monitoring results, realize accurate fault prediction of communication rooms, improve fault prediction efficiency, and contribute to the safe management and stable operation of communication rooms.

[0006] In a first aspect, an embodiment of the present application provides a communication room monitoring method based on big data, including:

[0007] Obtain the environment time series data, equipment operation time series data and inspection frequency time series data of the communication room in the historical time period;

[0008] After preprocessing the environmental time series data, equipment operation time series data, and inspection frequency time series data, the pre-trained prediction model is input to obtain the estimated environmental data, estimated equipment operation data, and estimated inspection frequency data of the communication room in the future set time period;

[0009] The prediction model is obtained by training the dynamic graph neural network based on causal perception through the historical environment time series data samples of the communication room, the historical equipment operation time series data samples and the historical inspection number time series data;

[0010] The estimated environmental data, estimated equipment operation data and estimated inspection frequency data are integrated to build a safety data set;

[0011] Evaluate the data priority of the security data set to obtain a priority evaluation result, and perform sub-packaging processing of the security data set based on the priority evaluation result to establish a sub-packaging data set;

[0012] The sub-package data sets are sorted in order according to the priority evaluation results, and a node status feature library is established. The sorted sub-package data sets and the node status feature library are input into the fault prediction model to perform fault prediction.

[0013] Among them, establishing a node status feature library includes:

[0014] Obtain node attribute information of the node, and construct node extreme performance constraints based on the node attribute information;

[0015] Reading the sending log of the node, performing continuous work analysis on the node based on the sending log, and establishing node fatigue, specifically including: obtaining the calling frequency of the node based on the sending log, and establishing a first fatigue parameter based on the calling frequency;

[0016] The node performance intensity is evaluated based on the sending log and the node limit performance constraint, and the second fatigue parameter is established according to the evaluation result and the current time series distance;

[0017] constructing a node fatigue according to a first fatigue parameter and a second fatigue parameter;

[0018] Node status evaluation is performed based on node extreme performance constraints and node fatigue to build a node status feature library.

[0019] The method further comprises: recording the fault prediction results, and generating node feedback data according to the recorded results;

[0020] The node status feature library is updated based on the node feedback data, and subsequent fault prediction is performed based on the node status database after feedback update.

[0021] The method pre-trains the prediction model in the following manner:

[0022] Obtain historical environment time series data samples, historical equipment operation time series data samples and historical inspection times time series samples of the communication room;

[0023] After the historical environment time series data samples, the historical equipment operation time series data samples and the historical inspection number time series samples are preprocessed, the historical environment time series data samples, the historical equipment operation time series data samples and the historical inspection number time series samples are converted into environment supervised data, equipment operation supervised data and inspection number supervised data respectively;

[0024] The supervised data of the environment, the supervised data of the equipment operation and the supervised data of the number of inspections are input into the dynamic graph neural network based on causal perception for training, and the prediction model is obtained after reaching the preset training cutoff condition.

[0025] The causal-aware dynamic graph neural network includes a sequentially connected input layer, a graph representation learning layer, a causal reasoning layer, a dynamic update layer and an output layer.

[0026] After the fault prediction is performed, the method further includes establishing first fault evaluation data, including:

[0027] Preset evaluation indicators;

[0028] Execute fault prediction and generate prediction results;

[0029] Based on the prediction results and the evaluation index, first fault evaluation data is established.

[0030] The method further comprises:

[0031] Based on the estimated environmental data, the estimated equipment operation data and the estimated inspection frequency data, the second fault evaluation data is established.

[0032] Comprehensive evaluation data is generated based on the first failure evaluation data and the second failure evaluation data.

[0033] The causal-aware dynamic graph neural network includes a sequentially connected input layer, a graph representation learning layer, a causal reasoning layer, a dynamic update layer and an output layer.

[0034] After the execution fault prediction, the method further includes:

[0035] If the communication room is predicted to be in a fault state, the state of the communication room is monitored in real time to obtain the communication room state monitoring information;

[0036] Perform abnormal monitoring based on the communication room status monitoring information.

[0037] The abnormality monitoring according to the communication room status monitoring information specifically includes:

[0038] Based on big data retrieval, abnormal instances of communication rooms are obtained, and the state change characteristics, time characteristics and location characteristics of abnormal phenomena of communication room abnormal instances are extracted, and then the abnormal knowledge graph is constructed by association.

[0039] Build an anomaly prediction model, construct a state space based on the anomaly knowledge graph, define the state change state of the abnormal phenomenon according to the state change trend of the abnormal instance in the communication room, and calculate the state transition probability through the state change characteristics of the abnormal instance in the communication room;

[0040] Construct a transition probability matrix based on the state transition probability, modify the transition probability matrix, set the initial state probability distribution, build an anomaly prediction model based on the state space, and build a training data set through the anomaly knowledge graph for model training;

[0041] Acquire the communication room status monitoring information, input the communication room status monitoring information into the abnormality prediction model for prediction analysis, and obtain abnormality prediction information;

[0042] Abnormal warning is carried out based on abnormal prediction information, warning information is generated, and feedback is given to management personnel.

[0043] In a second aspect, the embodiment of the present application provides a communication room monitoring system based on big data, including:

[0044] The first acquisition module is used to obtain the environment time series data, equipment operation time series data and inspection number time series data of the communication room in the historical time period;

[0045] The prediction module is used to pre-process the environment time series data, equipment operation time series data and inspection frequency time series data, and input the pre-trained prediction model to obtain the estimated environment data, estimated equipment operation data and estimated inspection frequency data of the communication room in the future set time period;

[0046] The prediction model is obtained by training the dynamic graph neural network based on causal perception through the historical environment time series data samples of the communication room, the historical equipment operation time series data samples and the historical inspection number time series data;

[0047] A fusion module is used to fuse the estimated environmental data, the estimated equipment operation data, and the estimated inspection frequency data, and construct a safety data set;

[0048] A first evaluation module is used to evaluate the data priority of the security data set, obtain a priority evaluation result, and perform sub-packaging processing of the security data set based on the priority evaluation result to establish a sub-packaging data set;

[0049] A fault prediction module is used to sort the sub-package data set in order according to the priority evaluation result, and to establish a node status feature library, and to input the sorted sub-package data set and the node status feature library into the fault prediction model to perform fault prediction;

[0050] Among them, establishing a node status feature library includes:

[0051] A second acquisition module is used to acquire node attribute information of the node and construct a node limit performance constraint based on the node attribute information;

[0052] The first establishment module is used to read the sending log of the node, perform continuous work analysis on the node based on the sending log, and establish node fatigue, specifically including: obtaining the calling frequency of the node based on the sending log, and establishing a first fatigue parameter based on the calling frequency;

[0053] The second establishment module is used to evaluate the node usage performance intensity based on the sending log and the node limit performance constraint, and establish a second fatigue parameter according to the evaluation result and the current time series distance;

[0054] A construction module, configured to construct a node fatigue according to a first fatigue parameter and a second fatigue parameter;

[0055] The second evaluation module is used to evaluate the node status based on the node limit performance constraint and node fatigue to build a node status feature library.

[0056] Compared with the prior art, the communication room monitoring method and system based on big data provided by the present application obtain the environment time series data, equipment operation time series data and inspection frequency time series data of the communication room in the historical time period; after pre-processing the environment time series data, equipment operation time series data and inspection frequency time series data, input the pre-trained prediction model to obtain the estimated environment data, estimated equipment operation data and estimated inspection frequency data of the communication room in the future set time period; the prediction model is obtained by training the dynamic graph neural network based on causal perception with the historical environment time series data samples, historical equipment operation time series data samples and historical inspection frequency time series data of the communication room; the estimated environment data, estimated equipment operation data and estimated inspection frequency data are fused, and a security data set is constructed; the data priority of the security data set is evaluated to obtain the priority evaluation result, and the security data set is sub-packaged based on the priority evaluation result to establish a sub-packaged data set; the sub-packaged data set is sorted according to the priority evaluation result, and a node status feature library is established, and the sorted sub-packaged data set and the node status feature library are input into the fault prediction model to execute Fault prediction; wherein, establishing a node status feature library includes: obtaining node attribute information of the node, and constructing a node extreme performance constraint based on the node attribute information; reading the node sending log, performing continuous work analysis on the node based on the sending log, and establishing node fatigue, specifically including: obtaining the call frequency of the node based on the sending log, and establishing a first fatigue parameter based on the call frequency; evaluating the node usage performance intensity based on the sending log and the node extreme performance constraint, and establishing a second fatigue parameter based on the evaluation result and the current time series distance; constructing node fatigue based on the first fatigue parameter and the second fatigue parameter; evaluating the node status based on the node extreme performance constraint and node fatigue to construct a node status feature library; through the above steps, not only the accuracy of data prediction is improved, the correlation and causal relationship between data are ensured, and the final monitoring results are made more stable and interpretable, but also the above method can make full use of the key information in the data to realize accurate fault prediction for future time periods, and improve the operation and maintenance efficiency and stability of the communication room; at the same time, the analysis of relevant data is strengthened to ensure that important data is processed in a timely manner, reduce the amount of fault prediction model calculations, and improve execution efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0058] Figure 1 A flow chart of a communication room monitoring method based on big data provided by the present application is shown;

[0059] Figure 2 A schematic diagram of a communication room monitoring system based on big data provided by the present application is shown;

[0060] Figure 3 A schematic diagram of an electronic device provided by the present application is shown. DETAILED DESCRIPTION

[0061] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0062] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by technicians in the field to which this application belongs.

[0063] In addition, the terms "first" and "second" etc. are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices.

[0064] Please refer to Figure 1 , Figure 1 A communication room monitoring method based on big data provided by an embodiment of the present application includes the following steps:

[0065] S101, obtaining the environment time series data, equipment operation time series data and inspection number time series data of the communication room in the historical time period;

[0066] In step S101, the environmental data includes temperature data, humidity data, air quality, and water accumulation in the communication room; specifically, the temperature can be collected by a temperature sensor, the humidity can be collected by a humidity sensor, the air quality can be obtained by a gas detector, and the water accumulation can be obtained by a water detector;

[0067] Equipment operation data includes communication records, user information, power parameters, and energy consumption management data;

[0068] The inspection frequency data refers to the number of times staff inspect the communication room;

[0069] In the historical time period, the above environmental data, equipment operation data and inspection frequency data are all counted as time series data, forming environmental time series data, equipment operation time series data and inspection frequency time series data respectively;

[0070] S102, after pre-processing the environment time series data, the equipment operation time series data and the inspection frequency time series data, inputting the pre-trained prediction model, obtaining the estimated environment data, the estimated equipment operation data and the estimated inspection frequency data of the communication room in the future set time period;

[0071] The prediction model is obtained by training the dynamic graph neural network based on causal perception through the historical environment time series data samples of the communication room, the historical equipment operation time series data samples and the historical inspection number time series data;

[0072] In this application, the method described herein pre-trains the prediction model in the following manner:

[0073] Obtain historical environment time series data samples, historical equipment operation time series data samples and historical inspection times time series samples of the communication room;

[0074] After the historical environment time series data samples, the historical equipment operation time series data samples and the historical inspection number time series samples are preprocessed, the historical environment time series data samples, the historical equipment operation time series data samples and the historical inspection number time series samples are converted into environment supervised data, equipment operation supervised data and inspection number supervised data respectively;

[0075] The supervised data of the environment, the supervised data of the equipment operation and the supervised data of the number of inspections are input into the dynamic graph neural network based on causal perception for training, and the prediction model is obtained after reaching the preset training cutoff condition.

[0076] The above-mentioned causal-aware dynamic graph neural network combines the graph structure data processing capability of the graph neural network and the deep interpretation capability of causal reasoning to process dynamically changing graph data. The dynamic graph neural network specifically includes a sequentially connected input layer, a graph representation learning layer, a causal reasoning layer, a dynamic update layer, and an output layer. The details are as follows:

[0077] Input layer: responsible for receiving the original dynamic graph data, which may include node features, edge information, and timestamps, etc., to describe the state of elements in the graph and the interactions between them;

[0078] Graph representation learning layer: This layer uses a graph neural network to encode the input graph data and extract feature representations of nodes and edges. These representations will capture the local and global structural information in the graph and provide a basis for subsequent processing;

[0079] Causal reasoning layer: In this layer, the model will introduce a causal reasoning mechanism, which can be achieved in many ways, such as building a structural causal model, using a causal discovery algorithm to identify causal relationships in the graph, or incorporating causal constraints into the training process of the graph neural network. The goal of this layer is to identify the causal dependencies between variables in the graph and use this to provide a deeper interpretation and analysis of the graph data.

[0080] Dynamic update layer: In view of the characteristics of dynamic graph data, this layer is responsible for processing the changes of elements in the graph over time; this may include updating the feature representation of nodes, adding or deleting edges, and processing time series data. By introducing the time dimension, the model can capture the dynamic changes in the graph and make predictions or decisions based on it;

[0081] Output layer: The output layer generates the final output based on the processing results of the previous layers. This may include the classification labels of nodes, the prediction results of edges, the attribute estimation of the graph, etc. The specific form of the output layer depends on the application scenario and task requirements of the model.

[0082] S103, fusing the estimated environmental data, the estimated equipment operation data, and the estimated inspection frequency data, and constructing a safety data set; specifically, this may be done by weighted fusion or the like;

[0083] S104, evaluating the data priority of the security data set to obtain a priority evaluation result, and performing sub-packaging processing of the security data set based on the priority evaluation result to establish a sub-packaging data set;

[0084] S105. Sort the sub-package data set in order according to the priority evaluation result, and establish a node status feature library, and input the sorted sub-package data set and the node status feature library into the fault prediction model to perform fault prediction; the fault prediction model is a tool that uses data and algorithms to predict possible faults or abnormal conditions in the communication room, and deep learning models such as convolutional neural network (CNN), recurrent neural network (RNN) and its variants (such as LSTM, GRU) can be selected;

[0085] Among them, establishing a node status feature library includes:

[0086] S1051, obtaining node attribute information of the node, and constructing a node limit performance constraint based on the node attribute information;

[0087] S1052, reading the sending log of the node, performing continuous work analysis on the node based on the sending log, and establishing node fatigue, specifically including: obtaining the calling frequency of the node based on the sending log, and establishing a first fatigue parameter based on the calling frequency;

[0088] S1053, performing node usage performance intensity evaluation based on the sending log and the node limit performance constraint, and establishing a second fatigue parameter according to the evaluation result and the current time series distance;

[0089] S1054, constructing node fatigue according to the first fatigue parameter and the second fatigue parameter;

[0090] S1055. Evaluate the node status based on the node limit performance constraint and node fatigue to build a node status feature library.

[0091] In this embodiment, the method further includes: recording the fault prediction results, and generating node feedback data according to the recorded results; the fault prediction results may include different states such as normal, warning or fault, and then the result of each prediction, the prediction time, the prediction object (such as node, component, etc.) and related information (such as environmental conditions, operation mode, etc.) are recorded in the database; the node feedback data may include prediction accuracy indicators (such as correct rate, false alarm rate, missed alarm rate, etc.), analysis of the causes of prediction errors, and any additional information that may affect the accuracy of the prediction (such as newly emerging abnormal modes, configuration changes, etc.);

[0092] The node status feature library is updated with feedback (such as adding new features, removing irrelevant features, adjusting feature weights or importance, etc.) through node feedback data, and subsequent fault prediction is performed based on the node status database after feedback update.

[0093] Through the above feedback and optimization mechanism, this application can continuously improve the accuracy and reliability of the fault prediction system, thereby more effectively ensuring the stable operation of the system.

[0094] In some embodiments of the present application, after performing the fault prediction, the method further includes establishing first fault evaluation data, including the following steps:

[0095] Preset evaluation indicators; these indicators may include but are not limited to accuracy, recall, F1 score, and false positive rate;

[0096] Perform fault prediction and generate prediction results; the prediction results may include the prediction category (such as normal, warning, fault) of each data sample and the probability or confidence of the prediction;

[0097] Based on the prediction results and evaluation indicators, the first fault evaluation data is established by weighted calculation, which can be used as a reference for staff. Establishing the first fault evaluation data by weighted calculation is a systematic and comprehensive process, which helps to evaluate and predict more accurately.

[0098] In some embodiments of the present application, the method further includes:

[0099] Based on the estimated environmental data, the estimated equipment operation data and the estimated inspection frequency data, and by assigning a weight to each data, the second fault evaluation data is established through weighted calculation.

[0100] Based on the first fault evaluation data and the second fault evaluation data, comprehensive evaluation data is generated through an existing weighted fusion method (such as a weighted average method), which helps to more accurately evaluate the health status and risk level of the communication room.

[0101] In some embodiments of the present application, after the execution of fault prediction, the method further includes:

[0102] If the communication room is predicted to be in a fault state, the state of the communication room is monitored in real time to obtain the communication room state monitoring information;

[0103] Carry out abnormal monitoring based on the communication room status monitoring information, and continue to strengthen monitoring.

[0104] In the above steps, the abnormality monitoring according to the communication room status monitoring information specifically includes:

[0105] Based on big data retrieval, abnormal instances of communication rooms are obtained, and the state change characteristics, time characteristics and location characteristics of abnormal phenomena of communication room abnormal instances are extracted, and then the abnormal knowledge graph is constructed by association.

[0106] Build an anomaly prediction model, construct a state space based on the anomaly knowledge graph, define the state change state of the abnormal phenomenon according to the state change trend of the abnormal instance in the communication room, and calculate the state transition probability through the state change characteristics of the abnormal instance in the communication room;

[0107] Construct a transition probability matrix based on the state transition probability, modify the transition probability matrix, set the initial state probability distribution, build an anomaly prediction model based on the state space, and build a training data set through the anomaly knowledge graph for model training;

[0108] Acquire the communication room status monitoring information, input the communication room status monitoring information into the abnormality prediction model for prediction analysis, and obtain abnormality prediction information;

[0109] Abnormal warning is carried out based on abnormal prediction information, warning information is generated, and feedback is given to management personnel.

[0110] The communication room monitoring method and system based on big data provided by the present application obtain the environment time series data, equipment operation time series data and inspection frequency time series data of the communication room in the historical time period; after pre-processing the environment time series data, equipment operation time series data and inspection frequency time series data, input the pre-trained prediction model to obtain the estimated environment data, estimated equipment operation data and estimated inspection frequency data of the communication room in the future set time period; the prediction model is obtained by training the dynamic graph neural network based on causal perception through the historical environment time series data samples, historical equipment operation time series data samples and historical inspection frequency time series data of the communication room; the estimated environment data, estimated equipment operation data and estimated inspection frequency data are fused, and a security data set is constructed; the data priority of the security data set is evaluated to obtain the priority evaluation result, and the security data set is sub-packed based on the priority evaluation result to establish a sub-packed data set; the sub-packed data set is sorted according to the priority evaluation result, and a node status feature library is established, and the sorted sub-packed data set and the node status feature library are input into the fault prediction model to perform fault prediction ; Among them, establishing a node status feature library includes: obtaining node attribute information of the node, and constructing a node extreme performance constraint based on the node attribute information; reading the node's sending log, performing continuous work analysis on the node based on the sending log, and establishing node fatigue, specifically including: obtaining the node's call frequency based on the sending log, and establishing a first fatigue parameter based on the call frequency; evaluating the node's usage performance intensity based on the sending log and the node extreme performance constraint, and establishing a second fatigue parameter based on the evaluation result and the current time series distance; constructing node fatigue based on the first fatigue parameter and the second fatigue parameter; evaluating the node status based on the node extreme performance constraint and node fatigue to construct a node status feature library; The above steps not only improve the accuracy of data prediction, ensure the correlation and causal relationship between data, and make the final monitoring results more stable and explainable, but also make full use of the key information in the data through the above method, realize accurate fault prediction for future time periods, and improve the operation and maintenance efficiency and stability of the communication room; At the same time, the analysis of relevant data is strengthened to ensure that important data is processed in a timely manner, reduce the amount of fault prediction model calculations, and improve execution efficiency.

[0111] In the above-mentioned embodiment, a communication room monitoring method based on big data is provided. Correspondingly, the present application also provides a communication room monitoring system based on big data. Figure 2As shown, the communication room monitoring system based on big data provided in the embodiment of the present application can implement the communication room monitoring method based on big data, and the communication room monitoring system based on big data can be implemented by software, hardware, or a combination of software and hardware. For example, the communication room monitoring system based on big data can include integrated or separate functional modules or units to perform the corresponding steps in the above methods, including:

[0112] The first acquisition module 101 is used to acquire the environment time series data, equipment operation time series data and inspection number time series data of the communication room in a historical time period;

[0113] Prediction module 102, used to pre-process the environment time series data, equipment operation time series data and inspection frequency time series data, and input the pre-trained prediction model to obtain the estimated environment data, estimated equipment operation data and estimated inspection frequency data of the communication room in the future set time period;

[0114] The prediction model is obtained by training the dynamic graph neural network based on causal perception through the historical environment time series data samples of the communication room, the historical equipment operation time series data samples and the historical inspection number time series data;

[0115] A fusion module 103 is used to fuse the estimated environmental data, the estimated equipment operation data and the estimated inspection number data, and construct a safety data set;

[0116] The first evaluation module 104 is used to evaluate the data priority of the security data set, obtain a priority evaluation result, and perform sub-packaging processing of the security data set based on the priority evaluation result to establish a sub-packaging data set;

[0117] The fault prediction module 105 is used to sort the sub-package data set according to the priority evaluation result, and establish a node status feature library, input the sorted sub-package data set and the node status feature library into the fault prediction model, and perform fault prediction;

[0118] Among them, establishing a node status feature library includes:

[0119] A second acquisition module is used to acquire node attribute information of the node and construct a node limit performance constraint based on the node attribute information;

[0120] The first establishment module is used to read the sending log of the node, perform continuous work analysis on the node based on the sending log, and establish node fatigue, specifically including: obtaining the calling frequency of the node based on the sending log, and establishing a first fatigue parameter based on the calling frequency;

[0121] The second establishment module is used to evaluate the node usage performance intensity based on the sending log and the node limit performance constraint, and establish a second fatigue parameter according to the evaluation result and the current time series distance;

[0122] A construction module, configured to construct a node fatigue according to a first fatigue parameter and a second fatigue parameter;

[0123] The second evaluation module is used to evaluate the node status based on the node limit performance constraint and node fatigue to build a node status feature library.

[0124] The communication room monitoring system based on big data provided in the embodiment of the present application and the communication room monitoring method based on big data provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, operated or implemented therein.

[0125] The embodiment of the present application also provides an electronic device corresponding to the method provided in the above embodiment, and the electronic device can be an electronic device used for a client, such as a mobile phone, a laptop computer, a tablet computer, a desktop computer, etc., to execute the above method.

[0126] Please refer to Figure 3 , which shows a schematic diagram of an electronic device provided by some embodiments of the present application. Figure 3 As shown, the electronic device 20 includes: a processor 200, a memory 201, a bus 202 and a communication interface 203, and the processor 200, the communication interface 203 and the memory 201 are connected via the bus 202; the memory 201 stores a computer program that can be run on the processor 200, and when the processor 200 runs the computer program, it executes the phishing email tracing method provided in any of the aforementioned embodiments of the present application.

[0127] The memory 201 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 203 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0128] The bus 202 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 201 is used to store programs, and the processor 200 executes the programs after receiving execution instructions. The method disclosed in any implementation of the above-mentioned embodiment of the present application may be applied to the processor 200, or implemented by the processor 200.

[0129] The processor 200 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 200. The above processor 200 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a readily available programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 201, and the processor 200 reads the information in the memory 201 and completes the steps of the above method in combination with its hardware.

[0130] The electronic device provided in the embodiment of the present application and the method provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the method adopted, operated or implemented by them.

[0131] The present application also provides a computer-readable storage medium corresponding to the method provided in the above-mentioned embodiment, on which a computer program (ie, a program product) is stored. When the computer program is executed by a processor, the method provided in any of the above-mentioned embodiments is executed.

[0132] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0133] The computer-readable storage medium provided in the above-mentioned embodiments of the present application and the method provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the method adopted, run or implemented by the application program stored therein.

[0134] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and they should all be included in the scope of the claims and specification of the present application.

Claims

1. A communication room monitoring method based on big data, characterized in that: The following steps are involved: Obtain the environment time series data, equipment operation time series data and inspection frequency time series data of the communication room in the historical time period; After preprocessing the environmental time series data, equipment operation time series data, and inspection frequency time series data, the pre-trained prediction model is input to obtain the estimated environmental data, estimated equipment operation data, and estimated inspection frequency data of the communication room in the future set time period; The prediction model is obtained by training the dynamic graph neural network based on causal perception through the historical environment time series data samples of the communication room, the historical equipment operation time series data samples and the historical inspection number time series data; The estimated environmental data, estimated equipment operation data and estimated inspection frequency data are integrated to build a safety data set; Evaluate the data priority of the security data set to obtain a priority evaluation result, and perform sub-packaging processing of the security data set based on the priority evaluation result to establish a sub-packaging data set; The sub-package data sets are sorted in order according to the priority evaluation results, and a node status feature library is established. The sorted sub-package data sets and the node status feature library are input into the fault prediction model to perform fault prediction. Among them, establishing a node status feature library includes: Obtain node attribute information of the node, and construct node extreme performance constraints based on the node attribute information; Reading the sending log of the node, performing continuous work analysis on the node based on the sending log, and establishing node fatigue, specifically including: obtaining the calling frequency of the node based on the sending log, and establishing a first fatigue parameter based on the calling frequency; The node performance intensity is evaluated based on the sending log and the node limit performance constraint, and the second fatigue parameter is established according to the evaluation result and the current time series distance; constructing a node fatigue according to a first fatigue parameter and a second fatigue parameter; Node status evaluation is performed based on node limit performance constraints and node fatigue to build a node status feature library; The method described herein pre-trains the prediction model in the following manner: Obtain historical environment time series data samples, historical equipment operation time series data samples and historical inspection times time series samples of the communication room; After the historical environment time series data samples, the historical equipment operation time series data samples and the historical inspection number time series samples are preprocessed, the historical environment time series data samples, the historical equipment operation time series data samples and the historical inspection number time series samples are converted into environment supervised data, equipment operation supervised data and inspection number supervised data respectively; The supervised data of the environment, the supervised data of the equipment operation and the supervised data of the number of inspections are input into the dynamic graph neural network based on causal perception for training, and the prediction model is obtained after reaching the preset training cutoff condition.

2. The method according to claim 1, characterized in that The method further comprises: recording the fault prediction results, and generating node feedback data according to the recorded results; The node status feature library is updated based on the node feedback data, and subsequent fault prediction is performed based on the node status database after feedback update.

3. The method according to claim 1, characterized in that The causal-aware dynamic graph neural network includes a sequentially connected input layer, a graph representation learning layer, a causal reasoning layer, a dynamic update layer and an output layer.

4. The method according to claim 1, characterized in that: After the fault prediction is performed, the method further includes establishing first fault evaluation data, including: Preset evaluation indicators; Execute fault prediction and generate prediction results; Based on the prediction results and the evaluation index, first fault evaluation data is established.

5. The method according to claim 4, characterized in that The method further comprises: Establishing second fault evaluation data based on the estimated environmental data, the estimated equipment operation data, and the estimated inspection frequency data; Comprehensive evaluation data is generated based on the first failure evaluation data and the second failure evaluation data.

6. The method according to claim 1, characterized in that After the execution fault prediction, the method further includes: If the communication room is predicted to be in a fault state, the state of the communication room is monitored in real time to obtain the communication room state monitoring information; Perform abnormal monitoring based on the communication room status monitoring information.

7. The method according to claim 6, characterized in that The abnormality monitoring according to the communication room status monitoring information specifically includes: Based on big data retrieval, abnormal instances of communication rooms are obtained, and the state change characteristics, time characteristics and location characteristics of abnormal phenomena of communication room abnormal instances are extracted, and then the abnormal knowledge graph is constructed by association. Build an anomaly prediction model, construct a state space based on the anomaly knowledge graph, define the state change state of the abnormal phenomenon according to the state change trend of the abnormal instance in the communication room, and calculate the state transition probability through the state change characteristics of the abnormal instance in the communication room; Construct a transition probability matrix based on the state transition probability, modify the transition probability matrix, set the initial state probability distribution, build an anomaly prediction model based on the state space, and build a training data set through the anomaly knowledge graph for model training; Acquire the communication room status monitoring information, input the communication room status monitoring information into the abnormality prediction model for prediction analysis, and obtain abnormality prediction information; Abnormal warning is carried out based on abnormal prediction information, warning information is generated, and feedback is given to management personnel.

8. The communication room monitoring system based on big data is characterized by: include: The first acquisition module is used to obtain the environment time series data, equipment operation time series data and inspection number time series data of the communication room in the historical time period; The prediction module is used to pre-process the environment time series data, equipment operation time series data and inspection frequency time series data, and input the pre-trained prediction model to obtain the estimated environment data, estimated equipment operation data and estimated inspection frequency data of the communication room in the future set time period; The prediction model is obtained by training the dynamic graph neural network based on causal perception through the historical environment time series data samples of the communication room, the historical equipment operation time series data samples and the historical inspection number time series data; A fusion module is used to fuse the estimated environmental data, the estimated equipment operation data, and the estimated inspection frequency data, and construct a safety data set; A first evaluation module is used to evaluate the data priority of the security data set, obtain a priority evaluation result, and perform sub-packaging processing of the security data set based on the priority evaluation result to establish a sub-packaging data set; A fault prediction module is used to sort the sub-package data set in order according to the priority evaluation result, and to establish a node status feature library, and to input the sorted sub-package data set and the node status feature library into the fault prediction model to perform fault prediction; Among them, establishing a node status feature library includes: A second acquisition module is used to acquire node attribute information of the node and construct a node limit performance constraint based on the node attribute information; The first establishment module is used to read the sending log of the node, perform continuous work analysis on the node based on the sending log, and establish node fatigue, specifically including: obtaining the calling frequency of the node based on the sending log, and establishing a first fatigue parameter based on the calling frequency; The second establishment module is used to evaluate the node usage performance intensity based on the sending log and the node limit performance constraint, and establish a second fatigue parameter according to the evaluation result and the current time series distance; A construction module, configured to construct a node fatigue according to a first fatigue parameter and a second fatigue parameter; The second evaluation module is used to evaluate the node status based on the node limit performance constraint and node fatigue to build a node status feature library; The prediction module pre-trains the prediction model in the following manner: Obtain historical environment time series data samples, historical equipment operation time series data samples and historical inspection times time series samples of the communication room; After the historical environment time series data samples, the historical equipment operation time series data samples and the historical inspection number time series samples are preprocessed, the historical environment time series data samples, the historical equipment operation time series data samples and the historical inspection number time series samples are converted into environment supervised data, equipment operation supervised data and inspection number supervised data respectively; The supervised data of the environment, the supervised data of the equipment operation and the supervised data of the number of inspections are input into the dynamic graph neural network based on causal perception for training, and the prediction model is obtained after reaching the preset training cutoff condition.

Images