Incremental clustering method, system, device and medium for classification of marine telegraph

Abstract

The present application relates to the field of message distribution, and provides a method, system, device and medium for classifying and querying marine telegrams, comprising: constructing a message element vector according to a message parameter; obtaining a message arrival interval of the marine telegrams, calculating a message arrival parameter according to the message arrival interval, obtaining a distribution mutation index of the classified level, and obtaining a receiving window according to the message arrival parameter and the distribution mutation index; receiving the marine telegrams in the receiving window, calculating a title similarity, a classified level asymmetric penalty factor, a number proximity, a level penalty factor and a priority penalty factor, and determining a classified level hard constraint; calculating a classified level constraint clustering distance, obtaining a received telegram cluster through the classified level constraint clustering distance and the classified level hard constraint; obtaining a query vector according to a query condition, obtaining a cluster abstract of the received telegram cluster, obtaining a candidate cluster according to the cluster abstract and the query vector, and selecting a query result from the candidate cluster.

Description

Technical Field

[0001] This invention relates to the field of message distribution technology, and in particular to a method, system, device and medium for incremental clustering of confidentiality-constrained marine telegraphs. Background Technology

[0002] Data stream clustering is a core research area in data mining. Existing data stream clustering methods can be mainly divided into four categories.

[0003] The first category is based on grid density, represented by PDStream and PAD-Stream. These methods divide the data space into grid cells, maintain grid density statistics online, and merge dense grids offline to extract clustering results. The drawback is that the number of grid cells in high-dimensional sparse data grows exponentially.

[0004] The second category is incremental clustering methods based on micro-clusters, represented by CluStream, DenStream, and ICDC. These methods maintain micro-cluster summaries online and have high incremental update efficiency. The drawback is that the micro-cluster summaries only contain numerical statistics and do not carry business semantics.

[0005] The third category is based on fuzzy granularity and adaptive windows, represented by AFGBStream. This type of method illustrates that the sliding window + adaptive decay method has become a common technical approach for data stream clustering, but its window adjustment is only based on time decay and does not perceive structural changes in business attributes.

[0006] The fourth category comprises methods for cluster merging and outlier handling, represented by incremental cluster merging methods based on recursive scalable aggregation functions. These methods demonstrate that incremental cluster merging and promoting new clusters from outlier buffers are mature techniques, but their merging decisions are based solely on statistical similarity and do not include compatibility checks for security semantics such as business confidentiality levels.

[0007] The above-mentioned existing technologies have four common shortcomings: First, there is currently no method to incorporate the business elements of the message into the clustering distance metric. Existing security classification methods follow a supervised classification approach, rather than using security level as a hard constraint and penalty factor for unsupervised clustering distance. Second, the sliding window is either fixed, statically determined by the elbow rule, or only performs general time decay, failing to detect sudden changes in security level or immediately narrow the window when such changes occur. Third, cluster merging decisions are based solely on statistical similarity or aggregation function values. When two clusters are statistically similar but have different business security levels, they are incorrectly merged, violating security isolation requirements. Fourth, existing methods stop at obtaining clustering results and do not support fuzzy queries based on business conditions. Summary of the Invention

[0008] This invention aims to at least solve one of the technical problems existing in related technologies. To this end, this invention provides a method, system, device, and medium for incremental clustering of marine telegrams based on their security classification, enabling the clustering of marine telegrams according to their security classification.

[0009] This invention provides a confidentiality-constrained incremental clustering method for marine telegraphs, comprising: S1: Generate marine telegrams, obtain the message parameters of marine telegrams, and construct message element vectors based on the message parameters; S2: Obtain the message arrival interval of the marine telegraph, calculate the message arrival parameters based on the message arrival interval, obtain the information level distribution change index, and obtain the receiving window based on the message arrival parameters and information level distribution change index. S3: Receive shipboard telegrams in the receiving window, calculate the title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor and priority penalty factor based on the message element vector, and determine the security level hard constraint. S4: Calculate the security level constraint clustering distance based on title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor, and priority penalty factor. Obtain the received telegram clusters through the security level constraint clustering distance and security level hard constraints. S5: Obtain the query conditions, obtain the query vector based on the query conditions, obtain the cluster summary of the received telegram cluster, obtain the candidate clusters based on the cluster summary and the query vector, select the query results from the candidate clusters, and complete the clustering and query of marine telegrams.

[0010] According to the confidentiality-constrained incremental clustering method for marine telegraphs provided by the present invention, step S1 further includes: S11: Generate a marine telegram and obtain message parameters from the marine telegram, including message number, message title, security classification, level, and priority. S12: Perform text conversion on the message title, retain the message number, and perform parameter mapping on the security level, grade, and priority to obtain the message element vector.

[0011] According to the confidentiality-constrained incremental clustering method for marine telegraphs provided by the present invention, step S2 further includes: S21: Obtain the message arrival interval of the marine telegraph, and calculate the message arrival parameters including arrival density and average arrival interval based on the message arrival interval; S22: Perform security classification statistics on the ship's telegrams to obtain the security classification distribution mutation index, determine the contraction and expansion conditions, and obtain the receiving window based on the message arrival parameters, the security classification distribution mutation index, the contraction and expansion conditions.

[0012] According to the incremental clustering method for marine telegraphs with security level constraints provided by the present invention, in step S3, a security level difference threshold is determined. The security level hard constraint is that when the security level difference between marine telegraphs is greater than the security level difference threshold, the security level constraint clustering distance is positive infinity.

[0013] According to the confidentiality-constrained incremental clustering method for marine telegraphs provided by the present invention, in step S4, the α-th marine telegraph... and the βth ship telegraph Clustering distance between density constraints The calculation method is as follows: in, As the first weighting coefficient, This is the second weighting coefficient. For title similarity, For the similarity of the numbers, It is a security level asymmetric penalty factor. As a graded penalty factor, This is a priority penalty factor.

[0014] According to the confidentiality-constrained incremental clustering method for marine telegraphs provided by the present invention, in step S4, historical telegraph data is obtained, a clustering distance threshold is obtained based on the confidentiality-constrained clustering distance of the historical telegraph data, and the received telegraph cluster is obtained through the confidentiality-constrained clustering distance, the confidentiality hard constraint, and the clustering distance threshold.

[0015] According to the confidentiality-constrained incremental clustering method for marine telegrams provided by the present invention, in step S5, the cluster keyword set and cluster distribution master pattern of the received telegram cluster are obtained, the cluster digest is obtained according to the cluster keyword set and the cluster distribution master pattern, the matching degree is calculated by query vector and cluster digest, and the candidate cluster is obtained by the matching degree.

[0016] This invention also provides a security-constrained incremental clustering system for marine telegraphs, comprising: Message element vector module: used to generate marine telegrams, obtain the message parameters of marine telegrams, and construct message element vectors based on the message parameters; The receiving window module is used to obtain the message arrival interval of marine telegrams, calculate the message arrival parameters based on the message arrival interval, obtain the density level distribution change index, and obtain the receiving window based on the message arrival parameters and density level distribution change index. Clustering constraint module: used to receive ship telegrams within the receiving window, calculate header similarity, security level asymmetric penalty factor, number proximity, grade penalty factor and priority penalty factor based on message element vectors, and determine security level hard constraints; Telegram Reception and Clustering Module: Used to calculate the security level constraint clustering distance based on header similarity, security level asymmetric penalty factor, number proximity, grade penalty factor, and priority penalty factor, and obtain the received telegram clusters through security level constraint clustering distance and security level hard constraints; Telegraph query module: used to obtain query conditions, obtain query vectors based on query conditions, obtain cluster digests of received telegraph clusters, obtain candidate clusters based on cluster digests and query vectors, select query results from candidate clusters, and complete the clustering and querying of marine telegraphs.

[0017] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the confidentiality-constrained incremental clustering method for marine telegraphs as described above.

[0018] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the confidentiality-constrained incremental clustering method for marine telegraphs as described above.

[0019] The above-described one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects: The present invention provides a method, system, device, and medium for incremental clustering of marine telegrams with security level constraints. It dynamically adjusts the receiving window by combining message arrival intervals and security level distribution abrupt change indicators. Under hard security level constraints, it calculates clustering distance by comprehensively considering header similarity, security level asymmetric penalty, number proximity, grade penalty factor, and priority penalty factor. This enables the accurate formation of received telegram clusters that meet security requirements, effectively avoiding erroneous aggregation of high-security and low-security telegrams, and enhancing the security and timeliness of clustering results. Simultaneously, the candidate cluster screening mechanism based on cluster summaries and query vectors significantly improves retrieval efficiency, thereby achieving controlled security level management, incremental clustering, and rapid querying of marine telegrams.

[0020] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating the confidentiality-constrained incremental clustering method for marine telegraphs provided by the present invention.

[0023] Figure 2 This is a schematic diagram of the structure of the confidentiality-constrained incremental clustering system for marine telegraphs provided by the present invention.

[0024] Figure 3 This is a schematic diagram of the structure of the confidentiality-constrained incremental clustering device for marine telegraphs provided by the present invention.

[0025] Figure label: 100. Message element vector module; 200. Receive window module; 300. Clustering constraint module; 400. Telegram receiving and clustering module; 500. Telegram query module; 810. Processor; 820. Communication interface; 830. Memory; 840. Communication bus. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. The following embodiments are used to illustrate this invention but cannot be used to limit the scope of this invention.

[0027] In the description of the embodiments of the present invention, it should be noted that the terms "first", "second" and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0028] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0029] The following is combined with Figures 1 to 3 Specific embodiments of the present invention are described below. Figure 1 This is a flowchart illustrating the confidentiality-constrained incremental clustering method for marine telegraphs provided by the present invention, including: S1: Generate marine telegrams, obtain the message parameters of marine telegrams, and construct message element vectors based on the message parameters; Furthermore, the objective of this stage is to construct a message element vector. Specifically, step S1 further includes: S11: Generate a marine telegram and obtain message parameters from the marine telegram, including message number, message title, security classification, level, and priority. S12: Perform text conversion on the message title, retain the message number, and perform parameter mapping on the security level, grade, and priority to obtain the message element vector.

[0030] First, a marine telegram needs to be generated. Then, message parameters, including message number, message title, security classification, level, and priority, are extracted from the business fields of the marine telegram header. Next, the message title is converted into a term frequency vector using the term frequency-inverse file frequency method. Security classification, level, and priority are then mapped. After mapping, for security classification, public is mapped to 0, internal to 1, secret to 2, confidential to 3, and top secret to 4; for level, ordinary is mapped to 0, urgent to 1, and urgent to 2; for priority, low priority is mapped to 0, medium priority to 1, and high priority to 2. The message number remains unchanged. Arranging the mapped values ​​of the message number, term frequency vector, security classification, level, and priority sequentially yields the message element vector.

[0031] S2: Obtain the message arrival interval of the marine telegraph, calculate the message arrival parameters based on the message arrival interval, obtain the information level distribution change index, and obtain the receiving window based on the message arrival parameters and information level distribution change index. Furthermore, the objective of this stage is to obtain the receiving window based on the message arrival parameters and the density level distribution abrupt change index. Specifically, step S2 further includes: S21: Obtain the message arrival interval of the marine telegraph, and calculate the message arrival parameters including arrival density and average arrival interval based on the message arrival interval; S22: Perform security classification statistics on the ship's telegrams to obtain the security classification distribution mutation index, determine the contraction and expansion conditions, and obtain the receiving window based on the message arrival parameters, the security classification distribution mutation index, the contraction and expansion conditions.

[0032] The specific implementation method for the above steps in this embodiment is as follows: The arrival process of marine telegraph messages is not uniform. During missions, a large number of urgent and priority messages arrive suddenly; while during routine duty, message arrivals are sparse. More importantly, the classification distribution changes drastically with mission status—routine duty is dominated by low-class messages, while during missions, a large number of high-class messages suddenly flood in. If the receiving window size is fixed, if the window is too small, clustering will fail due to insufficient message quantity; if the window is too large, a large number of old messages will interfere with clustering. Even more seriously, if the classification distribution changes abruptly and the window does not shrink, the newly arriving high-class messages will be diluted by a large number of historical low-class messages, resulting in the inability to achieve classification isolation of the received telegraph clusters.

[0033] Therefore, it is necessary to obtain the reception time of the marine telegraph messages. Based on this reception time, the time interval between telegraph arrivals can be determined, i.e., the message arrival interval. The mean of the message arrival intervals can then be calculated to obtain the arrival interval average. The reciprocal of the mean arrival interval average can be used as the arrival density, which can only be a maximum of 50 messages / second. This yields the message arrival parameters, including both the arrival density and the mean arrival interval.

[0034] Subsequently, based on the message element vector of the marine telegram, the proportion of each security level in the previous receiving window is calculated. It is also necessary to read the proportion of each security level in the receiving windows prior to the previous one. The change in the proportion of each security level between the two receiving windows is calculated, and the maximum change in the proportion of the security level is used as the security level distribution abrupt change index κ. Then, the contraction conditions are determined, including the following: The density-level distribution mutation index is higher than the preset density-level distribution mutation index threshold, which is set to 0.2 here. The arrival density is greater than the preset upper limit of arrival density, which is 2 lines / second.

[0035] When any one of the shrinkage conditions is met, the length of the previous receiving window is halved to obtain the current receiving window.

[0036] The expansion conditions also need to be determined, which include the following: The arrival density is less than the preset lower limit of arrival density, which is 0.2 lines / second. The density-level distribution mutation index is less than or equal to the preset density-level distribution mutation index threshold.

[0037] When the expansion condition is fully met, the previous receiving window is expanded by half, thus obtaining the current receiving window. Here, the receiving window must not be smaller than the preset minimum receiving window value.

[0038] S3: Receive shipboard telegrams in the receiving window, calculate the title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor and priority penalty factor based on the message element vector, and determine the security level hard constraint. Furthermore, the objective of this stage is to calculate the title similarity, security classification asymmetric penalty factor, number proximity, grade penalty factor, and priority penalty factor, and to determine the security classification hard constraint. Specifically, in step S3, a security classification difference threshold is determined. The security classification hard constraint is that when the security classification difference between ship telegrams is greater than the security classification difference threshold, the security classification constraint clustering distance is positive infinity.

[0039] The specific implementation method for the above steps in this embodiment is as follows: Subsequently, the recipient of the marine telegram receives the telegram within the time specified in the receiving window. Then, for each received telegram, the α-th telegram is calculated. and the βth ship telegraph Title similarity between : in, Let α be the word frequency vector of the α-th marine telegram. Let be the word frequency vector of the β-th marine telegram. To obtain the Euclidean norm.

[0040] Next, the asymmetric penalty factor for the security level between the α-th and β-th marine telegrams is calculated. Specifically, when the security classifications of the two ship telegrams are equal, the security classification asymmetric penalty factor is 0. When the security classification of the α-th ship telegram is higher than that of the β-th ship telegram, the value of the security classification asymmetric penalty factor is: Where c is the first penalty coefficient and d is the second penalty coefficient. The mapping of the security classification of the α-th marine telegram. This is the mapping of the security classification of the β-th ship telegram.

[0041] When the security classification of the α-th marine telegram is lower than that of the β-th marine telegram, the value of the security classification asymmetric penalty factor is: Next, the similarity in serial numbers between the α-th and β-th marine telegrams is calculated. : in, Let α be the message number of the αth marine telegram. This is the message number of the βth marine telegram.

[0042] The rank penalty factor between the α-th and β-th marine telegrams was then calculated. : Where γ is the level penalty coefficient, For the mapping of the α-th marine telegraph level, This is the mapping of the βth ship telegraph level.

[0043] Finally, the priority penalty factor between the α-th and β-th marine telegrams is calculated. : in, This is the priority penalty coefficient. The priority mapping for the α-th marine telegraph. This is the priority mapping for the β-th marine telegraph.

[0044] In addition, a hard constraint on the security classification needs to be established. Here, the hard constraint on the security classification means that when the difference in security classification between ship telegrams is greater than a preset security classification difference threshold, the clustering distance of the security classification constraint is directly set to positive infinity. In this embodiment, the security classification difference threshold is set to 1. The security classification difference here is the difference in the mapping between the security classifications of ship telegrams.

[0045] S4: Calculate the security level constraint clustering distance based on title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor, and priority penalty factor. Obtain the received telegram clusters through the security level constraint clustering distance and security level hard constraints. Furthermore, the objective of this stage is to calculate the density-constrained clustering distance, and to obtain the received telegraph cluster through the density-constrained clustering distance and the density-constrained hard constraints. Specifically, in step S4, the α-th marine telegraph... and the βth ship telegraph Clustering distance between density constraints The calculation method is as follows: in, As the first weighting coefficient, This is the second weighting coefficient. For title similarity, For the similarity of the numbers, It is a security level asymmetric penalty factor. As a graded penalty factor, This is a priority penalty factor.

[0046] In step S4, historical telegram data is obtained, and a clustering distance threshold is obtained based on the density-level constraint clustering distance of the historical telegram data. The received telegram cluster is obtained through the density-level constraint clustering distance, the density-level hard constraint, and the clustering distance threshold.

[0047] The specific implementation method for the above steps in this embodiment is as follows: Here, we first need to calculate the α-th marine telegraph. and the βth ship telegraph Clustering distance between density constraints : in, As the first weighting coefficient, This is the second weighting coefficient. For title similarity, For the similarity of the numbers, It is a security level asymmetric penalty factor. As a graded penalty factor, This is a priority penalty factor.

[0048] Next, it is necessary to determine the clustering distance threshold ε. Here, the security level constraint clustering distance of the ship's telegrams in the previous time window can be obtained as historical telegram data. The 80th percentile of all security level constraint clustering distances in the previous time window is then used as the clustering distance threshold. Subsequently, the security level constraint clustering distance is adjusted through hard security level constraints, and the received ship's telegrams with security level constraint clustering distances less than the clustering distance threshold are clustered to obtain initial clusters. For each cluster, this embodiment needs to record its cluster identifier, the number of messages within the cluster, the most frequently occurring keywords within the cluster, the proportion of each security level in the cluster, the proportion of each priority level in the cluster, the proportion of each grade level in the cluster, the minimum and maximum values ​​of the message numbers within the cluster, and the cluster's creation and update timestamps. This information is used as the cluster summary information.

[0049] Next, as other marine telegrams are added within the time window, the average of the message numbers in each initial cluster can be used as the virtual message number. Within the initial clusters, the word frequency vectors of keywords with high frequency in the telegram titles are used to construct virtual word frequency vectors. The average of the security classification mappings in the initial clusters is used as the virtual security classification, the average of the grade mappings in the initial clusters is used as the virtual grade, and the average of the priority mappings in the initial clusters is used as the virtual priority. In this way, virtual messages for the initial clusters can be constructed using virtual priority, virtual security classification, virtual grade, virtual word frequency vectors, and virtual message numbers. The security classification constraint clustering distance between each newly added marine telegram and each virtual message is calculated. Marine telegrams can be added to the initial cluster with the smallest security classification constraint clustering distance that is less than the clustering distance threshold. If no suitable initial cluster can be found, the telegram is temporarily stored in the outlier buffer. Here, after calculating the security level constraint clustering distance between the ship's telegrams and each virtual message, it is necessary to adjust the security level constraint clustering distance using security level hard constraints. This process continues until the end of the time window, at which point the received telegram cluster is obtained. The number of ship's telegrams in the received telegram cluster must not be less than the preset minimum number of telegrams.

[0050] For outlier buffers, when the number of marine telegrams within the buffer exceeds twice the minimum number of telegrams, the marine telegrams can be clustered in the outlier buffer following the steps described above. If the number of marine telegrams in the resulting cluster is greater than or equal to the minimum number of telegrams, then the cluster can be considered as a receiving telegram cluster. If a marine telegram in the outlier buffer fails to be assigned to any receiving telegram cluster within multiple time windows, it is discarded.

[0051] When the clustering distance between the virtual messages of two receiving telegram clusters is less than the clustering distance threshold, the two receiving telegram clusters can be merged. Here, merging the receiving telegram clusters is only allowed when the Jaccard similarity of the virtual word frequency vectors between the two virtual messages is greater than 0.3, and the maximum value of the difference between the proportions of each security level in the two receiving telegram clusters is less than 0.5.

[0052] S5: Obtain the query conditions, obtain the query vector based on the query conditions, obtain the cluster summary of the received telegram cluster, obtain the candidate clusters based on the cluster summary and the query vector, select the query results from the candidate clusters, and complete the clustering and query of marine telegrams.

[0053] Furthermore, the objective of this stage is to obtain candidate clusters, select query results from the candidate clusters, and complete the clustering and querying of marine telegrams. Specifically, in step S5, the cluster keyword set and cluster distribution master pattern of the received telegram clusters are obtained, a cluster summary is obtained based on the cluster keyword set and cluster distribution master pattern, the matching degree is calculated through the query vector and the cluster summary, and candidate clusters are obtained through the matching degree.

[0054] The specific implementation method for the above steps in this embodiment is as follows: First, the query conditions need to be obtained, which are the query keywords, query number range, query security classification, query priority, and query level of the marine telegrams to be queried, entered by the queryer. This allows the mapping of query keywords, query number range, query security classification, query priority, and query level to form the query vector. The mapping rules here are the same as in step S1. Next, the cluster digest information of the received telegram clusters needs to be obtained. This allows the cluster keyword set and cluster distribution master pattern to be obtained from the cluster digest information. The cluster keyword set consists of the keywords that appear most frequently in the received telegram clusters; the cluster distribution master pattern includes the master priority, master security classification, and master level. Here, the master priority is the highest priority appearing in the received telegram cluster, the master security classification is the highest security classification appearing in the received telegram cluster, and the master level is the highest level appearing in the received telegram cluster. The cluster keyword set and the cluster distribution master pattern are used as the cluster digest.

[0055] This allows for the calculation of keyword matching degree, security level matching degree, grade matching degree, and priority matching degree. Here, the cosine similarity between the query keyword and the cluster keyword set is used as the keyword matching degree. The security level matching degree is 1 if the mapping of the query security level is the same as the primary security level, and 0 otherwise. Similarly, the grade matching degree is 1 if the mapping of the query grade is the same as the primary grade, and 0 otherwise. The priority matching degree is 1 if the mapping of the query priority is the same as the primary priority, and 0 otherwise. The matching degree is obtained by weighted summation of the keyword matching degree, security level matching degree, grade matching degree, and priority matching degree. Received telegram clusters with matching degrees higher than a preset matching degree threshold are selected as candidate clusters. Here, the weights for keyword matching degree and security level matching degree are 0.4, and the weights for the others are 0.1.

[0056] Finally, the median of the query number range in the query vector is used as the virtual query number, and the query number range is replaced with the virtual query number. The security constraint clustering distance between the query vector after replacing the query number range and the marine telegrams in the candidate cluster can be calculated. The marine telegram with the smallest security constraint clustering distance is recommended to the queryer first. This completes the clustering and querying of marine telegrams.

[0057] The following describes the confidentiality-constrained incremental clustering device for marine telegraphs provided by the present invention. The confidentiality-constrained incremental clustering device for marine telegraphs described below can be referred to in correspondence with the confidentiality-constrained incremental clustering method for marine telegraphs described above.

[0058] Figure 2 A schematic diagram of the structure of a security-constrained incremental clustering system for marine telegraphs is shown, such as... Figure 2As shown, the confidentiality-constrained incremental clustering method for performing marine telegrams as described above includes: Message element vector module 100: used to generate marine telegrams, obtain the message parameters of marine telegrams, and construct message element vectors based on the message parameters; The receiving window module 200 is used to obtain the message arrival interval of marine telegrams, calculate the message arrival parameters based on the message arrival interval, obtain the density level distribution change index, and obtain the receiving window based on the message arrival parameters and the density level distribution change index. Clustering constraint module 300: Used to receive ship telegrams within the receiving window, calculate the header similarity, security level asymmetric penalty factor, number proximity, grade penalty factor and priority penalty factor based on the message element vector, and determine the security level hard constraint; Telegram receiving and clustering module 400: used to calculate the security level constraint clustering distance based on header similarity, security level asymmetric penalty factor, number proximity, grade penalty factor and priority penalty factor, and obtain the received telegram clusters through security level constraint clustering distance and security level hard constraints; Telegraph query module 500: used to obtain query conditions, obtain query vectors based on query conditions, obtain cluster summaries of received telegraph clusters, obtain candidate clusters based on cluster summaries and query vectors, select query results from candidate clusters, and complete the clustering and querying of marine telegraphs.

[0059] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call a computer program in the memory 830 to execute a confidentiality-constrained incremental clustering method for marine telegraphs, the method including: S1: Generate marine telegrams, obtain the message parameters of marine telegrams, and construct message element vectors based on the message parameters; S2: Obtain the message arrival interval of the marine telegraph, calculate the message arrival parameters based on the message arrival interval, obtain the information level distribution change index, and obtain the receiving window based on the message arrival parameters and information level distribution change index. S3: Receive shipboard telegrams in the receiving window, calculate the title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor and priority penalty factor based on the message element vector, and determine the security level hard constraint. S4: Calculate the security level constraint clustering distance based on title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor, and priority penalty factor. Obtain the received telegram clusters through the security level constraint clustering distance and security level hard constraints. S5: Obtain the query conditions, obtain the query vector based on the query conditions, obtain the cluster summary of the received telegram cluster, obtain the candidate clusters based on the cluster summary and the query vector, select the query results from the candidate clusters, and complete the clustering and query of marine telegrams.

[0060] Furthermore, when the computer program in the aforementioned memory 830 can be implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0061] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described incremental clustering methods for confidentiality constraints of marine telegraphs, the method comprising: S1: Generate marine telegrams, obtain the message parameters of marine telegrams, and construct message element vectors based on the message parameters; S2: Obtain the message arrival interval of the marine telegraph, calculate the message arrival parameters based on the message arrival interval, obtain the information level distribution change index, and obtain the receiving window based on the message arrival parameters and information level distribution change index. S3: Receive shipboard telegrams in the receiving window, calculate the title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor and priority penalty factor based on the message element vector, and determine the security level hard constraint. S4: Calculate the security level constraint clustering distance based on title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor, and priority penalty factor. Obtain the received telegram clusters through the security level constraint clustering distance and security level hard constraints. S5: Obtain the query conditions, obtain the query vector based on the query conditions, obtain the cluster summary of the received telegram cluster, obtain the candidate clusters based on the cluster summary and the query vector, select the query results from the candidate clusters, and complete the clustering and query of marine telegrams.

[0062] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0063] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A confidentiality-constrained incremental clustering method for marine telegraphs, characterized in that, include: S1: Generate marine telegrams, obtain the message parameters of marine telegrams, and construct message element vectors based on the message parameters; S2: Obtain the message arrival interval of the marine telegraph, calculate the message arrival parameters based on the message arrival interval, obtain the information level distribution change index, and obtain the receiving window based on the message arrival parameters and information level distribution change index. S3: Receive shipboard telegrams in the receiving window, calculate the title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor and priority penalty factor based on the message element vector, and determine the security level hard constraint. S4: Calculate the security level constraint clustering distance based on title similarity, security level asymmetric penalty factor, number proximity, grade penalty factor, and priority penalty factor. Obtain the received telegram clusters through the security level constraint clustering distance and security level hard constraints. S5: Obtain the query conditions, obtain the query vector based on the query conditions, obtain the cluster summary of the received telegram cluster, obtain the candidate clusters based on the cluster summary and the query vector, select the query results from the candidate clusters, and complete the clustering and query of marine telegrams.

2. The confidentiality-constrained incremental clustering method for marine telegraphs according to claim 1, characterized in that, Step S1 further includes: S11: Generate a marine telegram and obtain message parameters from the marine telegram, including message number, message title, security classification, level, and priority. S12: Perform text conversion on the message title, retain the message number, and perform parameter mapping on the security level, grade, and priority to obtain the message element vector.

3. The confidentiality-constrained incremental clustering method for marine telegraphs according to claim 1, characterized in that, Step S2 further includes: S21: Obtain the message arrival interval of the marine telegraph, and calculate the message arrival parameters including arrival density and average arrival interval based on the message arrival interval; S22: Perform security classification statistics on the ship's telegrams to obtain the security classification distribution mutation index, determine the contraction and expansion conditions, and obtain the receiving window based on the message arrival parameters, the security classification distribution mutation index, the contraction and expansion conditions.

4. The confidentiality-constrained incremental clustering method for marine telegraphs according to claim 1, characterized in that, In step S3, a confidentiality level difference threshold is determined. The confidentiality level hard constraint is that when the confidentiality level difference between ship telegrams is greater than the confidentiality level difference threshold, the confidentiality level constraint clustering distance is positive infinity.

5. The confidentiality-constrained incremental clustering method for marine telegraphs according to claim 1, characterized in that, In step S4, the αth marine telegraph and the βth ship telegraph Clustering distance between density constraints The calculation method is as follows: in, As the first weighting coefficient, This is the second weighting coefficient. For title similarity, For the similarity of the numbers, It is a security level asymmetric penalty factor. As a graded penalty factor, This is a priority penalty factor.

6. The confidentiality-constrained incremental clustering method for marine telegraphs according to claim 1, characterized in that, In step S4, historical telegram data is obtained, and a clustering distance threshold is obtained based on the density-level constraint clustering distance of the historical telegram data. The received telegram cluster is obtained through the density-level constraint clustering distance, the density-level hard constraint, and the clustering distance threshold.

7. The confidentiality-constrained incremental clustering method for marine telegraphs according to claim 1, characterized in that, In step S5, the cluster keyword set and cluster distribution master pattern of the received telegram cluster are obtained, the cluster digest is obtained according to the cluster keyword set and cluster distribution master pattern, the matching degree is calculated by query vector and cluster digest, and the candidate cluster is obtained by the matching degree.

8. A confidentiality-constrained incremental clustering system for marine telegraphs, used to execute the confidentiality-constrained incremental clustering method for marine telegraphs as described in any one of claims 1 to 7, characterized in that, include: Message element vector module: used to generate marine telegrams, obtain the message parameters of marine telegrams, and construct message element vectors based on the message parameters; The receiving window module is used to obtain the message arrival interval of marine telegrams, calculate the message arrival parameters based on the message arrival interval, obtain the density level distribution change index, and obtain the receiving window based on the message arrival parameters and density level distribution change index. Clustering constraint module: used to receive ship telegrams within the receiving window, calculate header similarity, security level asymmetric penalty factor, number proximity, grade penalty factor and priority penalty factor based on message element vectors, and determine security level hard constraints; Telegram Reception and Clustering Module: Used to calculate the security level constraint clustering distance based on header similarity, security level asymmetric penalty factor, number proximity, grade penalty factor, and priority penalty factor, and obtain the received telegram clusters through security level constraint clustering distance and security level hard constraints; Telegraph query module: used to obtain query conditions, obtain query vectors based on query conditions, obtain cluster digests of received telegraph clusters, obtain candidate clusters based on cluster digests and query vectors, select query results from candidate clusters, and complete the clustering and querying of marine telegraphs.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the confidentiality-constrained incremental clustering method for marine telegraphs as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the confidentiality-constrained incremental clustering method for marine telegraphs as described in any one of claims 1 to 7.

Images