Knowledge system model, method and product for ship behavior analysis

Through the knowledge system model for ship behavior analysis, static and dynamic data are integrated, the problem of low supervision and analysis efficiency of multi-source spatio-temporal data at sea is solved, accurate analysis and dynamic update of ship behavior is achieved, and the accuracy of analysis results is improved. reliability.

CN118966334AActive Publication Date: 2024-11-15BEIJING BIG DATA ADVANCED TECH RES INST
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202410991639.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-15
Estimated Expiration
2044-07-23

AI Technical Summary

Technical Problem

Traditional manual duty methods are difficult to meet the regulatory needs of multi-source spatiotemporal data at sea. Massive data contain a large amount of useless noise information and low-quality data, resulting in low efficiency in ship behavior analysis and difficulty in obtaining accurate results.

Method used

A knowledge system model for ship behavior analysis was designed, including object knowledge layer, activity model layer and topic event layer. By integrating static ship information and real-time dynamic data, the behavior analysis results of the target ship are generated, and the ship activity law category is used. Knowledge and behavioral analysis models combine multi-source data for correlation analysis to provide accurate and comprehensive correlation data.

Benefits of technology

It realizes an accurate and comprehensive analysis of ship behavior, integrates dispersed data through logical hierarchical design and diversified associations, and discovers hidden associations, which improves the accuracy and reliability of analysis results, and supports dynamic updates and maintenance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118966334A_ABST
    Figure CN118966334A_ABST
Patent Text Reader

Abstract

The invention provides a ship behavior analysis-oriented knowledge system model, method and product, and relates to the technical field of ship behavior analysis, and the model comprises an object knowledge layer which is used as a knowledge model of a bottom layer, is used for integrating static ship information into entities, and is used for storing entity association relationships among the entities; the activity model layer is used for providing support for analysis of ship behaviors and comprises ship activity rule knowledge and a ship behavior analysis model; the ship behavior analysis model generates a behavior analysis result of the target ship in the occurrence process of the target theme event according to real-time dynamic data which is extracted from the theme event layer and is related to the target theme event by utilizing ship activity law class knowledge; and the subject event layer is used for storing multi-source real-time dynamic data in the form of sub-event nodes, each sub-event node belongs to a corresponding subject event, and displaying a behavior analysis result of the target ship to a service user.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of ship behavior analysis, and in particular to a knowledge system model, method and product for ship behavior analysis. Background Art

[0002] The multi-source spatiotemporal data of maritime traffic is growing rapidly, and the traditional manual on-duty method is difficult to meet the regulatory needs of maritime management. To achieve efficient supervision of various types of ships and their abnormal behaviors, it is necessary to conduct intelligent analysis of ship behavior. The rapid development of the Internet has led to the generation of massive amounts of data every day. These data contain a lot of information, which provides a basis for the automated and intelligent behavior analysis of ships. However, these data also contain a lot of useless noise information and low-quality data. Business personnel need to spend a lot of time and experience to screen and mine the data, making it difficult to achieve ideal results in ship behavior analysis.

[0003] Therefore, it is necessary to develop a knowledge system model, method and product for ship behavior analysis, so as to provide accurate and comprehensive correlation data for ship behavior analysis, and then generate more accurate ship behavior analysis results. Summary of the invention

[0004] In view of the above problems, the embodiments of the present application provide a knowledge system model, method and product for ship behavior analysis to overcome the above problems or at least partially solve the above problems.

[0005] In a first aspect of an embodiment of the present application, a knowledge system model for ship behavior analysis is provided, the knowledge system model comprising: an object knowledge layer, an activity model layer and a subject event layer;

[0006] The object knowledge layer is used as the underlying knowledge model to integrate static ship information into entities and store entity association relationships between entities. The entities include: ship entities, person entities, organization entities, account entities, time entities and location entities.

[0007] The activity model layer is used to provide support for the analysis of ship behavior, and includes: ship activity law knowledge and a ship behavior analysis model; the ship behavior analysis model uses the ship activity law knowledge to generate a behavior analysis result of the target ship during the occurrence of the target theme event based on the real-time dynamic data related to the target theme event extracted from the theme event layer;

[0008] The subject event layer is used to store real-time dynamic data from multiple sources in the form of sub-event nodes, wherein each sub-event node belongs to a corresponding subject event, and displays the behavior analysis results of the target ship to business users, or sub-events in various fields related to the target subject event that are connected in series according to time elements; wherein the real-time dynamic data includes: ship activity trajectory sub-event nodes, task activity sub-event nodes, account content sub-event nodes, news report sub-event nodes, and event association relationships between each sub-event node.

[0009] The second aspect of the embodiment of the present application further provides a ship behavior analysis method, which is applied to the knowledge system model for ship behavior analysis described in the first aspect of the embodiment of the present application, including:

[0010] Receive target subject event query information input by the owner user;

[0011] Input the target subject event query information into the ship behavior analysis model, use the ship activity law class knowledge, and generate the target ship behavior analysis result during the occurrence of the target subject event according to the real-time dynamic data related to the target subject event extracted from the subject event layer;

[0012] The behavior analysis result of the target ship is displayed to the business user, or the sub-events in various fields related to the target theme event are connected in series according to the time element.

[0013] The third aspect of an embodiment of the present application also provides an electronic device, including a memory, a processor, and a computer program stored on the memory, wherein the processor executes the computer program to implement the steps in a ship behavior analysis method described in the second aspect of an embodiment of the present application.

[0014] The fourth aspect of the embodiment of the present application further provides a computer-readable storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the steps in the ship behavior analysis method described in the second aspect of the embodiment of the present application are implemented.

[0015] The fifth aspect of the embodiment of the present application further provides a computer program product, which, when executed on an electronic device, enables a processor to implement the steps in a ship behavior analysis method as described in the second aspect of the embodiment of the present application.

[0016] The embodiment of the present application provides a knowledge system model for ship behavior analysis, the knowledge system model includes: an object knowledge layer, an activity model layer and a theme event layer; the object knowledge layer is used as a bottom knowledge model to integrate static ship information into entities and store entity association relationships between entities, the entities include: ship entities, person entities, organization entities, account entities, time entities and location entities; the activity model layer is used to provide support for the analysis of ship behavior, including: ship activity law class knowledge, and a ship behavior analysis model; the ship behavior analysis model uses the ship activity law class knowledge to extract the relevant information from the theme event layer The real-time dynamic data related to the target theme event generates the behavior analysis results of the target ship during the occurrence of the target theme event; the theme event layer is used to store real-time dynamic data from multiple sources in the form of sub-event nodes, wherein each sub-event node belongs to the corresponding theme event, and displays the behavior analysis results of the target ship to business users, or the sub-events in various fields related to the target theme event that are connected in series according to time elements; wherein the real-time dynamic data includes: ship activity trajectory sub-event nodes, task activity sub-event nodes, account content sub-event nodes, news report sub-event nodes, and event association relationships between each sub-event node.

[0017] The embodiment of the present application is oriented to the needs of ship behavior analysis, establishes a ship behavior analysis knowledge system model, standardizes the structure and association of multi-source data, and organizes all information related to business analysis in a logical and orderly manner, organizes and integrates different types of knowledge in a hierarchical manner, and forms a unified framework to adapt to the ever-changing environment and needs. The specific beneficial effects are: on the one hand, the embodiment of the present application defines a logically layered knowledge system model, modularly organizes and manages relevant knowledge and data, so that static background information (static ship information) and dynamic real-time data can be uniformly processed in a clear, orderly, and easy-to-understand manner. On the other hand, the embodiment of the present application integrates and connects the scattered data by constructing a diversified association (entity association relationship and event association relationship) of different data within and between levels, which helps to discover the hidden associations between data, show the development trend of related events, and verify the accuracy of the data, so that more accurate and comprehensive ship behavior related information can be obtained in practical applications. In addition, the embodiment of the present application further incorporates the analyzed laws (ship activity law knowledge) and algorithm models (ship behavior analysis models) into the knowledge system model for management, and synchronizes and updates, maintains and reuses them with the continuous addition of new data to generate more accurate ship behavior analysis results. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0019] Figure 1 It is a structural schematic diagram of a knowledge system model for ship behavior analysis provided by an embodiment of the present application;

[0020] Figure 2 It is a schematic diagram of a framework of entity association relationships of an object knowledge layer provided in an embodiment of the present application;

[0021] Figure 3 It is a flow chart of the steps of a ship behavior analysis method provided by an embodiment of the present application;

[0022] Figure 4 It is a schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] The exemplary embodiments of the present application will be described in more detail below in conjunction with the accompanying drawings in the embodiments of the present application. Although the exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present application and to enable the scope of the present application to be fully communicated to those skilled in the art.

[0024] The multi-source spatiotemporal data of maritime traffic is growing rapidly, and the traditional manual monitoring method is difficult to meet the supervision needs of maritime management. To achieve the supervision of various types of ships and their abnormal behaviors, it is necessary to conduct intelligent analysis of ship behavior and improve the situational awareness of ships in relevant sea areas by using multi-source spatiotemporal data, such as studying ship behavior trends based on track correlation and event analysis, predicting abnormal behavior of ships based on law learning and domain knowledge, and dynamically identifying ships based on the motion model and speed characteristics of ships.

[0025] The rapid development of the Internet has led to the generation of massive amounts of data every day. These data contain a large number of rich records of the occurrence and development of events, which provide a basis for the automated and intelligent behavior analysis of ships. But at the same time, these data also contain a large amount of useless noise information and low-quality data. Business personnel need to spend a lot of time and experience to screen and mine the data. The problem of information overload has seriously restricted the efficiency of data processing and the release of data value. At present, users generally use Internet search engines to obtain the required text, pictures, videos and other data, but most search engines are based on supply and demand matching and sorting algorithms to find relevant content, but cannot provide a full picture of events and development context that affect ship behavior. It is still difficult to manually integrate and sort multi-source information and build the context environment required for analysis. How to effectively and quickly transform data into valuable and explainable knowledge information has become the primary issue in current ship behavior analysis.

[0026] With the rapid development of natural language processing technology, some network platforms and search engines can now provide certain event context generation functions for some hot topics, and can provide time information and brief descriptions of the development process of related events. The information source of the event is mostly news reports on the platform, and does not cover information from other channels. However, in the era of mobile Internet, events that affect the movement of ships or information that reflects the current behavior of ships may not only appear in formal news reports, but also in network information that the public generally participates in discussions and in real-time activity messages of ships. Therefore, to establish an automated and intelligent ship behavior analysis process, it is necessary to combine historical data mining, explore the current status, behavior laws and spatiotemporal characteristics of ships. These feature-based knowledge cannot be directly obtained and guaranteed to be correct through a single source of information, and it is necessary to integrate multi-source data for correlation analysis, feature mining and cross-validation.

[0027] At present, the behavior analysis of ships has the problems of insufficient use of multi-source data and difficulty in establishing associations between data. Specifically: 1) Massive information is widely dispersed in various information spaces, with different carriers and complex data types, showing disorder and multi-dimensional characteristics. How to deeply mine fine-grained and multi-dimensional knowledge that meets the needs of ship behavior analysis from massive, multi-source, heterogeneous and dynamically changing data is the primary problem. Most of the existing solutions currently rely only on ship automatic identification system (Automatic Identification System, AIS) data for analysis. Data from a single source has the problems of limited information volume, insufficient data completeness, and difficulty in verifying accuracy, which may cause the analysis results to be out of touch with reality and have no practical application value. It is necessary to increase the data source, define the structure of multi-source data, and clarify data requirements, collection content and data format. 2) Ship behavior is easily affected by the environment and current events, which may be hidden in multi-source data. Therefore, it is necessary to build a logical relationship between data to show the development and evolution of events. The construction of multi-source data association requires mining and discovery based on the inherent attributes of events, entities, and various static and dynamic information. How to organize and screen various types of data in an orderly manner, extract the knowledge required for analysis, and then support the retrospective analysis of events that affect ship behavior is a key step in studying ship behavior patterns and predicting ship movements and anomalies. The trajectory-based analysis method does not take into account the impact of the event environment and current situation on ship behavior, which may cause the analysis results to be out of touch with reality and have low accuracy. It is necessary to combine dynamic information such as current news and ship trajectories with static information such as historical data and background knowledge, and incorporate ship behavior into real environment scenarios for analysis, proofreading and verification to improve the reliability and practicality of the analysis results.

[0028] In view of the above problems, the embodiments of the present application propose that it is necessary to develop a knowledge system model, method and product for ship behavior analysis, so as to provide accurate and comprehensive related data for ship behavior analysis, and then generate more accurate ship behavior analysis results. The knowledge system model for ship behavior analysis provided by the embodiments of the present application is described in detail below through some embodiments and their application scenarios in conjunction with the accompanying drawings.

[0029] The first aspect of the embodiment of the present application provides a knowledge system model for ship behavior analysis. Figure 1 , Figure 1 A schematic diagram of the structure of a knowledge system model for ship behavior analysis provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the knowledge system model includes: object knowledge layer, activity model layer and subject event layer;

[0030] The object knowledge layer is used as the underlying knowledge model to integrate static ship information into entities and store entity association relationships between entities. The entities include: ship entities, person entities, organization entities, account entities, time entities and location entities.

[0031] The activity model layer is used to provide support for the analysis of ship behavior, including: ship activity law knowledge and a ship behavior analysis model; the ship behavior analysis model uses the ship activity law knowledge to generate a behavior analysis result of the target ship during the occurrence of the target theme event based on the real-time dynamic data related to the target theme event extracted from the theme event layer.

[0032] The subject event layer is used to store real-time dynamic data from multiple sources in the form of sub-event nodes, wherein each sub-event node belongs to a corresponding subject event, and displays the behavior analysis results of the target ship to business users, or sub-events in various fields related to the target subject event that are connected in series according to time elements; wherein the real-time dynamic data includes: ship activity trajectory sub-event nodes, task activity sub-event nodes, account content sub-event nodes, news report sub-event nodes, and event association relationships between each sub-event node.

[0033] The knowledge system model constructed in the embodiment of the present application is as follows Figure 1As shown, from top to bottom are the subject event layer, the activity model layer and the object knowledge layer. The knowledge system can also be understood as a knowledge structure, which is a comprehensive system of various types of knowledge that is reorganized and constructed with a certain logic after the original data is cleaned and processed. The knowledge system covers multi-source and multi-domain data, and integrates scattered data into unified knowledge through logic and internal associations. The knowledge system is usually composed of concepts, associations, facts and other contents. In the scenario of ship behavior analysis proposed in this embodiment, the knowledge system is a summary of all knowledge related to the ship's status, behavior, and environment. These knowledge are summarized and organized to form a set of structured knowledge frameworks. Therefore, the knowledge system model models and represents the semantic concepts, characteristic attributes, logical associations, etc. of things related to ship behavior analysis, forming a network or tree-like hierarchical structure of domain knowledge, making the knowledge more systematic, structured and easy to understand, and used to guide business personnel to convert the underlying multi-source heterogeneous data into directly usable knowledge for the scenario. The knowledge system model proposed in this embodiment includes various open source raw data related to ships as collection objects, takes the historical behavior of ships and the development of events as background knowledge (i.e., entities in the object knowledge layer), is based on expert experience and reasoning logic (corresponding to the ship activity rules in the activity model layer), uses artificial intelligence natural language processing algorithms as tools (i.e., the ship behavior analysis model in the activity model layer), and uses human-machine combined association analysis as a means to identify, extract, and construct structured data features with business value, forming a knowledge system model for ship behavior analysis, opening up the cleaning, processing, and integration technology chains from data to information and from information to knowledge, and providing data support for further comprehensive analysis of ship behavior and determination of abnormal conditions.

[0034] Among them, the object knowledge layer, as the underlying knowledge model, classifies, integrates and aggregates static information (i.e., static ship information) according to entity concepts, stores it as multiple entities, and stores the entity association relationships between entities, thereby allowing users to query and access background knowledge that is important to the upper-level business in a unified and structured manner, so as to conduct accurate and in-depth scenario analysis and modeling based on these entity concepts, and provide basic support for the fusion and association of multi-source data. The object knowledge layer provides a large amount of static background information related to ships (i.e., static ship information that exists as background information, and static information refers to information that will not change easily within a certain period of time that is long enough), which provides an important basis for building the association between upper-level multi-source data (i.e., the event association relationship between each sub-event node in the subject event layer) and entity disambiguation. Specifically, the construction of the object knowledge layer includes modeling six entities of ships, people, organizations, accounts, time, and places in the network platform and their corresponding relationships, that is, the entities include: ship entities, person entities, organization entities, account entities, time entities, and place entities.

[0035] The ship entity is the most core entity in the object knowledge layer, which records the inherent attributes and activity attributes of a specific ship. The entity attribute information of the ship entity includes: ship name, ship type, ship model, ship attributes, hull number (the unique identifier of the ship) and typical speed range (that is, the commonly used speed range when the target is sailing, in m / s).

[0036] People related to ships can also provide important clues to the behavior logic of ships, such as dynamic information published by crew members in public channels, speeches or movements of relevant people, etc. The person entities in the object knowledge layer include: ship work person entities and important position person entities. Summarizing and sorting out the background information of the person is the basis for establishing the relationship between the ship and the account content and news. The entity attribute information of the person entity includes: the person's Chinese name, the person's English name, abbreviation, the person's region, position (such as captain, etc.), a brief description of the person, and the storage address of the person's related multimedia resources.

[0037] As an entity strongly associated with a person, the entity attribute information of the account entity includes: account ID, account username, user gender, user avatar, account platform and user homepage information, which is used to assist in establishing the association between the person and the account content (for example, the association between the person entity and the account content sub-event node).

[0038] In addition, for some shipping-related organizations, the events or related trends, announcements, etc. in which they participate are likely to have a significant impact on the behavior of ships. Therefore, the embodiment of the present application can assist in building the association between people, news, and ships by establishing an organization entity. The entity attributes of the organization entity include: the Chinese name of the organization, the English name of the organization, the region to which the organization belongs, the address of the organization, a brief description of the organization, and the storage address of the organization's multimedia resources.

[0039] Spatiotemporal information is an important basis for the theme event layer to establish the temporal relationship and geographical association relationship between sub-events (based on the spatiotemporal information, the sub-events in various fields related to the target theme event are connected in series according to the time element, thereby generating the behavior analysis results). However, the spatiotemporal feature granularity of multi-source heterogeneous data (i.e. the acquired raw data, static ship information) often differs. In order to support the fusion of spatiotemporal data, it is necessary to sort out the various dimensional information of time and place as entity attribute information in the object knowledge layer as comprehensively as possible, and obtain the corresponding attribute values ​​according to certain conversion rules, so as to align the spatiotemporal elements and enhance the flexibility of entity search. The entity attribute information of the time entity includes: timestamp, date, year, month, festival (i.e. the name of the festival corresponding to the current date), week, zodiac, lunar calendar, zodiac calendar and time description information (i.e. the fuzzy text description of time). The entity attribute information of the place entity includes: location longitude and latitude, region, address and place description information (i.e. the fuzzy text description of the place).

[0040] In a possible implementation, the object knowledge layer further includes: an entity association module, configured to generate the entity association relationship between entities according to entity attribute information of each entity.

[0041] Specifically, in the object knowledge layer, the entity association module can establish an association between entities based on entity attribute information. When it is determined that there is entity attribute information associated between two entities, the entity association relationship between the two entities (such as Figure 1 The arrows connecting the entities in the object knowledge layer shown in the figure represent the association relationship between the entities. For example, the association relationship between the person entity and the ship entity is determined by the person's position information (i.e., the entity attribute information of the position in the person entity) or the affiliation information, and the collaborative association relationship between the ship entities is determined by the affiliation information, the accompanying behavior, etc. Figure 2 , Figure 2 A schematic diagram of the entity association relationship of an object knowledge layer is shown, such as Figure 2 As shown, entity association relationships include: person relationships between person entities (including superior-subordinate relationships, colleague relationships, and account friend relationships), institutional relationships between institutional entities (including superior / subordinate institutions), position relationships between person entities and institutional entities, position relationships between person entities and ship entities (such as captains, crew members, and the ships they serve on, etc.), ownership relationships between person entities and account entities, affiliation relationships between ship entities and institutional entities, collaborative relationships between ship entities (including accompanying and escorting), time association relationships (including before, after, and in the same time period), and geographic association relationships (including located in and nearby areas). Considering that the collaborative association relationship between ship entities may change with the environment and time, in order to enhance flexibility and accuracy, the "effective time" attribute is defined for the collaborative association relationship between ship entities to indicate the time range corresponding to the collaborative association relationship. For example, by comparing the similarity of the trajectories of two ships within a certain time range, it can be determined that the two ships have an accompanying relationship within the time period.

[0042] In this embodiment, an activity model layer is provided above the object knowledge layer to provide support for the analysis of ship behavior, including: ship activity law knowledge and a ship behavior analysis model. The ship behavior analysis model uses the ship activity law knowledge to generate a behavior analysis result of the target ship during the occurrence of the target theme event based on the real-time dynamic data related to the target theme event extracted from the theme event layer.

[0043] Specifically, the activity model layer is used to store semi-static ship activity law knowledge obtained based on statistical analysis, as well as algorithm models for reasoning and analysis (i.e., ship behavior analysis models), to provide support for the analysis, evaluation, and prediction of ship behavior activities, and to provide external analysis and mining results of spatiotemporal data. Exemplarily, a user can select a target subject event a to obtain real-time dynamic data of all sub-events in the subject event layer that are connected in time sequence under the subject event a. The user can input the behavior analysis requirements for a specific ship A, and based on the real-time dynamic data of the target subject event a and the ship activity law knowledge of the target ship A, the extracted real-time dynamic data is analyzed using the ship behavior analysis model to obtain the behavior analysis results of the target ship A.

[0044] The ship activity law knowledge in the activity model layer mainly includes: ship activity area law, ship activity time law, ship activity route law and ship activity trend law, which are usually obtained based on statistics, manual annotation or data mining. The ship activity area law is based on the historical trajectory data of the ship, and is obtained by experts through statistics and analysis to obtain the area where the ship often moves. The law attribute information of the ship activity area law includes: activity area type (i.e. circular area, polygonal area and coded statistical area), activity area content, activity area location and specific description information of the ship activity area law. Among them, the activity area content includes: the central longitude, central latitude, area radius and area probability of the circular area; the area boundary and area probability of the polygonal area; the area statistics of the coded statistical area. Refer to the following Table 1, which shows a law attribute information of the activity area content.

[0045]

[0046] Table 1

[0047] The ship activity time pattern is a regular value obtained by statistical analysis based on expert experience or data mining methods according to different time scales, rather than a single time point. The regular attribute information of the ship activity time pattern includes: time statistical scale (such as monthly, daily, hourly, etc.), activity time content (describes the specific content of the activity time pattern, which can be expressed in json text, such as: {"time1": 0.1, "time2": 0.05, ...}), activity time area (describes the area or sea area where the ship is located during the activity time) and specific description information of the ship activity time pattern.

[0048] The ship activity route law describes the route that ships often travel, also known as the classic track. The regular attribute information of the ship activity route law includes: activity route content (describing the longitude and latitude information of the activity route, which can be expressed in a json text list, such as: [{"lon":108.5,"lat":37.65}, ...]), route starting area, route ending area, route passing area and specific description information of the ship activity route law.

[0049] The ship activity pattern records the frequency of ship activities between a specific starting point and end point. The pattern attribute information includes: departure point (latitude and longitude information expressed in JSON text), arrival point, number of activities and specific description information of the ship activity pattern.

[0050] In a possible implementation, the ship behavior analysis model includes: a ship behavior dynamic identification model, a ship movement prediction model, and a ship behavior abnormality warning model;

[0051] The knowledge system model also includes:

[0052] A training module is used to train the ship behavior analysis model based on a machine learning algorithm using the real-time dynamic data stored in the subject event layer.

[0053] In this embodiment, the algorithm model (ship behavior analysis model) in the activity model layer can be obtained based on the real-time data in the subject event layer and trained based on the machine learning algorithm (the specific type of the machine learning algorithm is not limited in this embodiment), including: ship behavior dynamic identification model, ship movement prediction model, and ship behavior abnormal warning model. Among them, the ship behavior dynamic identification model obtains the current behavior recognition result of the target ship based on the behavior information of the target ship in the current time period obtained in the subject event layer. The ship movement prediction model is used to predict the behavior of the target ship at the next moment based on the behavior information of the target ship in the current time period obtained in the subject event layer. The ship behavior abnormal warning model is based on the behavior information of the target ship in the current time period obtained in the subject event layer, combined with the stored knowledge of the relevant ship activity rules of the target ship, and analyzed to determine whether the target ship is currently in an abnormal state. Correspondingly, the ship behavior analysis results generated based on the above analysis model include at least one of the following: based on the real-time dynamic data of the current subject event, identify the ship type corresponding to the target trajectory (dynamic identification result of the ship behavior dynamic identification model), predict the next movement of the target ship (movement prediction result of the ship movement prediction model), and judge whether the current ship status is abnormal (abnormal warning result of the ship behavior abnormal warning model). Before actual application, the real-time dynamic data of the subject event layer can also be used as training data to train each ship behavior analysis model to improve the accuracy of model analysis.

[0054] In a possible implementation, the ship behavior analysis model is stored in the activity model layer as an independent entity;

[0055] The model attribute information of the ship behavior analysis model includes: model unique identification ID, model function name, model type, model file, input parameters, output parameters, creation time and model function body description.

[0056] Specifically, the optimized algorithm model (ship behavior analysis model) is stored as an independent entity. Users can understand the model content and call it directly through the key information stored in the entity (such as model type, model parameters and description, etc.), thereby simplifying the analysis process and enhancing the reusability of the model. The model attribute information of the ship behavior analysis model as an entity includes: model unique identification ID, model function name (name used to represent the model function), model type (such as rule model, deep learning model, Bayesian model, decision tree model, etc.), model file (i.e. binary serialization data of the model, containing relevant feature calculation functions), input parameters (json text describing the model input parameters and types), output parameters, creation time and specific description of the model function. As an intermediate layer, the activity model layer mainly defines the association relationship between the ship entity and the ship activity law class knowledge, decouples the ship behavior analysis model and the background knowledge base (object knowledge layer), and provides the upper layer (theme event layer) with the analysis results of different ship behavior laws by associating various entities in the object knowledge layer, so that the theme event layer can show the behavior analysis results of the target ship to the business users, making the analysis process clearer and more convenient.

[0057] Specifically, there are certain dependencies between different levels in the knowledge system model. The analysis of the ship behavior analysis model in the activity model layer needs to rely on the entity attribute information of the relevant entities provided by the object knowledge layer. For example, the ship dynamic recognition model needs to be supported by background data such as cruising speed and typical speed range (i.e., entity attribute information of the object knowledge layer) when performing recognition. At the same time, the establishment of various ship behavior analysis models also requires the combination of algorithms, rules and refined activity laws. For example, the abnormal warning of ships needs to rely on regular data such as their typical trajectories and activity ranges (ship activity law knowledge) for calculation.

[0058] In an embodiment of the present application, a subject event layer is provided above the model activity layer, which is used to store real-time dynamic data from multiple sources in the form of sub-event nodes, wherein each sub-event node belongs to a corresponding subject event, and displays the behavior analysis results of the target ship to business users, or the sub-events in various fields related to the target subject event that are connected in series according to time elements.

[0059] The reasons that lead to the corresponding behavior of the ship may be hidden in a series of multi-domain events that can be obtained through open source information. For example, emergencies may occur, which in turn affect the activity patterns of the ship and the psychological state of the relevant personnel, and leave relevant activity tracks and information. Therefore, ship behavior analysis not only relies on normalized background knowledge, but also has mutual correlation with emergency tasks, events, network information, etc. In the subject event layer, based on a certain subject scenario (i.e., a specific subject event), the multi-source information such as the ship's activity track, task execution, related network information and events can be organized and displayed in a timeline to verify whether various information can match each other, and by calling the relevant algorithm model (ship behavior analysis model) to conduct a retrospective analysis of the event, sort out the trend of the current event development and the environmental situation, so as to summarize the logic and intention of the ship's behavior and obtain the ship behavior analysis results. The embodiment of the present application proposes a knowledge system model for ship behavior analysis. From the perspective of information organization, it describes events, entities and their relationships by designing a logical hierarchical organization framework for spatiotemporal data, and constructs a scenario-responsive knowledge management model in a targeted manner to support the orderly organization of static and dynamic data, the automated extraction of knowledge, and the retrospective analysis of the development of related events, thereby providing intuitive, accurate, and comprehensive intelligence knowledge support for ship situation awareness.

[0060] In this embodiment, the real-time dynamic data includes: ship activity track sub-event nodes, task activity sub-event nodes, account content sub-event nodes, news report sub-event nodes, and event association relationships between each sub-event node. For each sub-event node of the subject event layer, the common event attribute information includes: node ID, subject and node type (i.e., enumeration values ​​corresponding to different sub-event nodes).

[0061] The ship activity trajectory sub-event node is used to describe the real-time position, speed and moving direction of the ship, which can be obtained after cleaning the data of the ship's automatic identification system (Automatic Identification System, AIS). The event attribute information of the ship activity trajectory sub-event node also includes: the unique identification ID of the ship's trajectory, reporting time, location longitude and latitude, location (that is, the area to which each track point in the ship's navigation trajectory belongs), speed, moving direction, and ship's hull number.

[0062] The mission activity sub-event node is used to describe the mission activities performed by the ship within a certain time and space range. It is a summary and refinement of the mission content and time and space characteristics after integrating multiple open source data. The event attribute information of the mission activity sub-event node also includes: target name, hull number, mission type (such as cargo transportation, escort, rescue, etc.), mission details, mission start time, mission end time, start location and end location (including location name and longitude and latitude).

[0063] The news report sub-event node is mainly used to record events related to ships. Reports published on online platforms and related institutional networks have the advantages of clear expression, high accuracy, large amount of information, strong objectivity, and complete temporal and spatial characteristics. They are the most important information source for the subject event layer. The event attribute information of the news report sub-event node includes: news title, event type, event time, event location (including location name and latitude and longitude), event subject, event object, event content, content summary, and multimedia resource link address attached to the event.

[0064] The account content sub-event node records the public expression information published by crew members and related people on the network platform. Although the account content may imply the current status and trends of the ship, it is usually mixed with the personal opinions of the characters, the quality is uncontrollable and has a certain degree of subjectivity. Therefore, it can only be used as an auxiliary clue to cross-verify the authenticity of other information. The account content sub-event focuses on extracting the temporal, spatial and content characteristics of the content. The event attribute information of the account content sub-event node includes: the link address of the account content post, the poster's account, the posting time, the posting location (including the location name and longitude and latitude), the link address of other resources (such as pictures, videos, etc.) attached to the account content post, and the time, space and content characteristics extracted from the account content text.

[0065] In a possible implementation, the subject event layer further includes:

[0066] The event association module is used to generate event association relations between each sub-event node according to the entity associated with each sub-event node and the object knowledge layer.

[0067] Specifically, within the subject event layer, there are certain associations between the sub-event nodes under a subject event, and the most common one is the temporal association (before, after, and same time period). In addition, different sub-event nodes can be linked through the object knowledge layer (i.e., determine the event association relationship).

[0068] The activity model layer also includes:

[0069] The rule association module is used to associate the ship activity rule knowledge with the entities of the object knowledge layer.

[0070] The rule model association module is used to generate the association relationship between the ship activity rule class knowledge and the ship behavior analysis model according to the entities of the object knowledge layer.

[0071] On the one hand, the activity law is based on the historical activity data of a certain target, obtained by expert statistics, manual annotation or data mining. Therefore, there is a natural association between the ship activity law knowledge and the entity of the object knowledge layer, and the ship activity law knowledge and the entity of the object knowledge layer can be associated through the law association module. Exemplarily, the ship activity law knowledge A is associated with the entity (ship a) of the object knowledge layer to illustrate that the ship activity law knowledge A is a knowledge record of the activity law of ship a. On the other hand, in the ship behavior analysis model, it is often necessary to rely on the activity law of the target in the activity model layer and the sub-event data related to the target in the subject event layer for analysis and judgment. Therefore, it is also necessary to generate the association relationship between the model algorithm (ship behavior analysis model) and the activity law (ship activity law knowledge) based on the target (i.e., the entity of the object knowledge layer) through the law model association module.

[0072] In a possible implementation manner, generating the event association relationship between the sub-event nodes according to the entities associated with the object knowledge layer includes:

[0073] Step S101, extracting event attribute information from the target sub-event node through deep learning and information extraction algorithms.

[0074] Specifically, for each selected sub-event node (taken as the target sub-event node), the subject event layer extracts event attribute information (which can be the attribute information related to the subject event in the event attribute information of the target sub-event node) based on deep learning and information extraction algorithms, such as news topic, event type (also understood as event trigger word), event subject, event object and time and space attributes.

[0075] Step S102: determining a target entity associated with the target sub-event node from the object knowledge layer according to the entity information in the event attribute information.

[0076] Specifically, the extracted event attribute information is classified and sorted to obtain the entity information associated therein (such as event subject, event object), so as to determine the target entity associated with the target sub-event node in the object knowledge layer. For example, for the news report sub-event node 1, the extracted event attribute information includes: event subject (ship A), then correspondingly, the entity information of ship A associated in the object knowledge layer can be determined and determined as the target entity (such as Figure 1 As shown, sub-event 1, sub-event 2, and sub-event 3 in the subject event layer are associated with the corresponding entities of the object knowledge layer such as "place", "account", "time", and "ship" through arrow connecting lines.

[0077] Step S103: determining one or more candidate entities associated with the target entity according to the entity association relationship of the target entity.

[0078] Specifically, the object knowledge layer stores various entities and the associations between entities, and one or more candidate entities that have entity associations with the target entity can be determined. For example, for the entity of ship A, the geographic association between ship A and region B can be found based on the existing entity associations, thereby determining the candidate entity region B.

[0079] Step S104, determining one or more candidate sub-event nodes associated with the target sub-event node, and the event association relationship between the target sub-event node and the candidate sub-event nodes, based on the sub-event nodes associated with the candidate entity.

[0080] Specifically, in accordance with steps S101 and S102, the sub-event node associated with the candidate entity is found (taking each sub-event node as the target sub-event node, the entity associated with each sub-event node in the object knowledge layer is determined respectively). Exemplarily, for the candidate entity determined according to region B, the associated sub-event node 1 is found, and then an event association relationship between news report sub-event node 1 and sub-event node 1 can be generated. The types of event association relationships between different sub-event nodes may be different, so it is also necessary to determine the type of the event association relationship based on the types of the two sub-event nodes, the associated entities (ship A, region B), and the entity association relationship involved (geographical association relationship between ship A and region B). Thus, by establishing an association between the target sub-event node and the object knowledge layer, the object knowledge layer is then used as a bridge to associate sub-event nodes of different types.

[0081] Specifically, the event association relationship between the news sub-event node and the account content sub-event node needs to be established by comprehensively considering the time relationship, the "news sub-event-ship" relationship, the "ship-person / institution" relationship, the "person / institution-account" relationship and the "account-account content sub-event" relationship; the event association relationship between the activity track sub-event node and the account content sub-event node needs to be established by comprehensively considering the time relationship, the "activity track sub-event-ship" relationship, the "ship-person / institution" relationship, the "person / institution-account" relationship and the "account-account content sub-event" relationship; the relationship between the news report sub-event node and the task activity sub-event node needs to be established by comprehensively considering the time relationship, the "news sub-event-ship" relationship and the "task activity sub-event-ship" relationship.

[0082] In the subject event layer, each sub-event node belongs to one or more corresponding subject events, and the sub-event nodes of the same subject event are classified under the same subject. Specifically, sub-event nodes from different platforms, trajectory data and other channels may have certain temporal and spatial associations, event associations, logical associations, causal associations, etc. Mining and sorting sub-events will help to better understand the occurrence patterns, development trends and potential risks of subject events, understand the specific impact of events on ship behavior on the basis of understanding the overall picture of the event, and provide factual basis for ship movement and abnormal prediction.

[0083] In a possible implementation, the theme event layer further includes: a theme event determination module, configured to determine the theme event corresponding to each event sub-node according to the following steps:

[0084] Step S201, extracting event semantic information from the target news report sub-event node through deep learning and information extraction algorithm

[0085] Step S202: determining the subject event to which the target news report sub-event node belongs according to the event semantic information.

[0086] Step S203, according to the event association relationship of the target news report sub-event node, determining the subject event to which the target news report sub-event node belongs as the subject event of the sub-event node associated with the target news report sub-event node.

[0087] Specifically, based on the news report sub-event node with rich semantic information, event semantic information is extracted from the news report sub-event node through deep learning and information extraction algorithms. The event semantic information represents the semantic information of the event involved in the sub-event node (in this embodiment, the deep learning method and information extraction algorithm used are not specifically limited). Then, based on the event semantic information, the subject event to which the news report sub-event node belongs is generated. Therefore, based on the association between the news sub-event node and other types of sub-event nodes, sub-event nodes that lack semantic information such as trajectories, account content, and tasks can be further associated with the subject event, which facilitates users to retrieve and discover clues related to a certain topic, making the information display at the subject layer more accurate and effective.

[0088] In this embodiment, the subject event layer can also show the behavior analysis results of the target ship to the business users, or the sub-events in various fields related to the target subject event connected in series according to the time element. The subject event layer is used to store real-time dynamic data from multiple sources on the one hand, and on the other hand, it combines knowledge at various levels to uniformly show the current situation and analysis results to business users. It can be understood as a scene-oriented information management framework connected in series according to the time element. The subject event layer is composed of a series of sub-event nodes around different subject events. These sub-events may occur independently or have certain dependencies or associations, but all sub-events are related to the subject event. However, due to the diverse data sources of this layer, the credibility and granularity of the data are different, and it has a certain dynamicity, it is difficult to directly integrate them with a unified algorithm. Therefore, a visual display method is selected to present the temporal association between the sub-events in the subject event layer, providing business personnel with the ability to find clues related to a certain subject event. As a scene-oriented information display layer, the subject event layer needs to call and integrate data (entity information of the object knowledge layer) and models (analysis models of the activity model layer) from the lower layers. Through the user interaction interface, it uniformly organizes and displays all contextual information and ship navigation tracks related to scene topics (i.e. specific subject events) within a certain time and space range for business personnel, forming an event context from the past to the present and the corresponding data evidence chain, realizing multi-perspective retrospection, analysis and cross-verification of event development.

[0089] In a possible implementation, the knowledge system model further includes:

[0090] The subject event layer updating module is used to update the sub-event nodes and event association relationships in the real-time dynamic data at intervals of a first time length.

[0091] The regularity knowledge updating module is used to update the regularity knowledge of the ship activities according to the real-time dynamic data.

[0092] In this embodiment, during the specific implementation, since the subject event layer stores real-time dynamic data, it is necessary to continuously obtain the latest data to update it. Some sub-event nodes in the subject event layer (for example, sub-event nodes with an occurrence date 3 months ago) can be replaced every first time interval. And based on the acquired new data, corresponding new sub-event nodes and corresponding event association relationships (including event association relationships between new sub-event nodes, and event association relationships between new sub-event nodes and previous sub-event nodes) are generated. The specific value of the first time length can be set according to the actual application situation. In this embodiment, it is not limited. With the continuous generation of new data, the ship activity law class knowledge of the activity model layer may also be further updated to ensure the accuracy of the relevant law knowledge.

[0093] The embodiment of the present application is oriented to the needs of ship behavior analysis. It combines background knowledge (entities in the object knowledge layer) and multi-source real-time data (sub-event nodes in the subject event layer) to establish a ship behavior analysis knowledge system model, aiming to standardize the structure and association of multi-source data, and organize all information related to business analysis in a logical and orderly manner. It organizes and integrates different types of knowledge in a hierarchical manner to form a unified framework to adapt to the ever-changing environment and needs, and realize "the combination of static and dynamic, and equal emphasis on data and logic."

[0094] The specific beneficial effect is that the knowledge system model proposed in this application realizes a logical hierarchical design. By defining a logically hierarchical knowledge system model, modularly organizing and managing relevant knowledge and data, static background information (i.e., entities in the object knowledge layer) and dynamic real-time data (i.e., sub-event nodes in the subject event layer) can be uniformly processed in a clear, orderly, and easy-to-understand manner; and scene-oriented (each subject event) modeling is realized. According to the needs of ship behavior analysis, the ontology structure design is carried out for data of different levels and types, which is the basis for feature extraction and association construction of data, and is the premise for transforming multi-source data into domain knowledge. Business analysis support is realized. On the one hand, by constructing diversified associations of different data within and between levels, the scattered data are merged and connected in series, which helps to discover hidden associations between data (i.e., entity association relationships and event association relationships), show the development trend of related events, and verify the accuracy of the data. On the other hand, the rules (ship activity rule knowledge) and algorithm models (ship behavior analysis models) obtained by analysis can be further incorporated into the knowledge system model for management, and synchronously updated, maintained and reused with the continuous addition of new data in the future.

[0095] The second aspect of the embodiment of the present application further provides a ship behavior analysis method, which is applied to the knowledge system model for ship behavior analysis described in the first aspect of the embodiment of the present application, referring to Figure 3 , Figure 3 A flow chart of the steps of a ship behavior analysis method is shown, such as Figure 3 As shown, the method includes:

[0096] Step S1, receiving target subject event query information input by an owner user.

[0097] Step S2, input the target subject event query information into the ship behavior analysis model, use the ship activity law knowledge, and generate the target ship behavior analysis results during the target subject event based on the real-time dynamic data related to the target subject event extracted from the subject event layer.

[0098] Step S3, showing the behavior analysis result of the target ship to the business user, or the sub-events in various fields related to the target theme event connected in series according to the time element.

[0099] The knowledge system model includes an object knowledge layer, an activity model layer and a subject event layer; the object knowledge layer is used as the underlying knowledge model to integrate static ship information into entities and store entity association relationships between entities, including ship entities, person entities, organization entities, account entities, time entities and location entities;

[0100] The activity model layer is used to provide support for the analysis of ship behavior, and includes: ship activity law knowledge and a ship behavior analysis model; the ship behavior analysis model uses the ship activity law knowledge to generate a behavior analysis result of the target ship during the occurrence of the target theme event based on the real-time dynamic data related to the target theme event extracted from the theme event layer;

[0101] The subject event layer is used to store real-time dynamic data from multiple sources in the form of sub-event nodes, wherein each sub-event node belongs to a corresponding subject event, and displays the behavior analysis results of the target ship to business users, or sub-events in various fields related to the target subject event that are connected in series according to time elements; wherein the real-time dynamic data includes: ship activity trajectory sub-event nodes, task activity sub-event nodes, account content sub-event nodes, news report sub-event nodes, and event association relationships between each sub-event node.

[0102] In a possible implementation, the object knowledge layer further includes: an entity association module, configured to generate the entity association relationship between entities according to entity attribute information of each entity;

[0103] The subject event layer further comprises: an event association module, which is used to generate event association relations between each sub-event node according to entities associated with each sub-event node and the object knowledge layer;

[0104] The activity model layer also includes:

[0105] A rule association module, used to associate the ship activity rule knowledge with the entities of the object knowledge layer;

[0106] The rule model association module is used to generate the association relationship between the ship activity rule class knowledge and the ship behavior analysis model according to the entities of the object knowledge layer.

[0107] In a possible implementation manner, generating the event association relationship between the sub-event nodes according to the entities associated with the object knowledge layer includes:

[0108] Through deep learning and information extraction algorithms, event attribute information is extracted from the target sub-event node;

[0109] Determining a target entity associated with the target sub-event node from the object knowledge layer according to the entity information in the event attribute information;

[0110] Determine one or more candidate entities associated with the target entity according to the entity association relationship of the target entity;

[0111] According to the sub-event nodes associated with the candidate entity, one or more candidate sub-event nodes associated with the target sub-event node and the event association relationship between the target sub-event node and the candidate sub-event nodes are determined.

[0112] In a possible implementation, the theme event layer further includes: a theme event determination module, configured to determine the theme event corresponding to each event sub-node according to the following steps:

[0113] Through deep learning and information extraction algorithms, event semantic information is extracted from the sub-event nodes of the target news report;

[0114] Determining the subject event to which the target news report sub-event node belongs according to the event semantic information;

[0115] According to the event association relationship of the target news report sub-event node, the subject event to which the target news report sub-event node belongs is determined as the subject event of the sub-event node associated with the target news report sub-event node.

[0116] In a possible implementation, the ship behavior analysis model includes: a ship behavior dynamic identification model, a ship movement prediction model, and a ship behavior abnormality warning model;

[0117] The knowledge system model also includes:

[0118] A training module is used to train the ship behavior analysis model based on a machine learning algorithm using the real-time dynamic data stored in the subject event layer.

[0119] In a possible implementation, the ship behavior analysis model is stored in the activity model layer as an independent entity;

[0120] The model attribute information of the ship behavior analysis model includes: model unique identification ID, model function name, model type, model file, input parameters, output parameters, creation time and model function body description.

[0121] In a possible implementation, the knowledge system model further includes:

[0122] A subject event layer updating module, used for updating the sub-event nodes and event association relationships in the real-time dynamic data at intervals of a first time length;

[0123] The regularity knowledge updating module is used to update the regularity knowledge of the ship activities according to the real-time dynamic data.

[0124] The present application also provides an electronic device, referring to Figure 4 , Figure 4 Schematic diagram of an electronic device proposed in an embodiment of the present application. Figure 4 As shown, the electronic device 100 includes: a memory 110 and a processor 120. The memory 110 and the processor 120 are connected via a bus communication. A computer program is stored in the memory 110. The computer program can be run on the processor 120 to implement the steps in the ship behavior analysis method disclosed in the embodiment of the present application.

[0125] The embodiment of the present application also provides a computer-readable storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the steps in the ship behavior analysis method disclosed in the embodiment of the present application are implemented.

[0126] The embodiment of the present application also provides a computer program product. When the computer program product is run on an electronic device, the processor implements the steps of the ship behavior analysis method disclosed in the embodiment of the present application.

[0127] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0128] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, apparatuses, electronic devices, and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0131] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present application.

[0132] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0133] The above is a detailed introduction to the knowledge system model, method and product for ship behavior analysis provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A knowledge system model for ship behavior analysis, It is characterized in that The knowledge system model includes: object knowledge layer, activity model layer and subject event layer; The object knowledge layer is used as the underlying knowledge model to integrate static ship information into entities and store entity association relationships between entities. The entities include: ship entities, person entities, organization entities, account entities, time entities and location entities. The activity model layer is used to provide support for the analysis of ship behavior, and includes: ship activity law knowledge and a ship behavior analysis model; the ship behavior analysis model uses the ship activity law knowledge to generate a behavior analysis result of the target ship during the occurrence of the target theme event based on the real-time dynamic data related to the target theme event extracted from the theme event layer; The subject event layer is used to store real-time dynamic data from multiple sources in the form of sub-event nodes, wherein each sub-event node belongs to a corresponding subject event, and displays the behavior analysis results of the target ship to business users, or sub-events in various fields related to the target subject event that are connected in series according to time elements; wherein the real-time dynamic data includes: ship activity trajectory sub-event nodes, task activity sub-event nodes, account content sub-event nodes, news report sub-event nodes, and event association relationships between each sub-event node.

2. The knowledge system model for ship behavior analysis according to claim 1, It is characterized in that The object knowledge layer further includes: an entity association module, which is used to generate the entity association relationship between entities according to entity attribute information of each entity; The subject event layer also includes: An event association module, used for generating event association relations between each sub-event node according to entities associated with each sub-event node and the object knowledge layer; The activity model layer also includes: A rule association module, used to associate the ship activity rule knowledge with the entities of the object knowledge layer; The rule model association module is used to generate the association relationship between the ship activity rule class knowledge and the ship behavior analysis model according to the entities of the object knowledge layer.

3. The knowledge system model for ship behavior analysis according to claim 2, It is characterized in that The generating of the event association relationship between the sub-event nodes according to the entities associated with the object knowledge layer includes: Through deep learning and information extraction algorithms, event attribute information is extracted from the target sub-event node; Determining a target entity associated with the target sub-event node from the object knowledge layer according to the entity information in the event attribute information; Determine one or more candidate entities associated with the target entity according to the entity association relationship of the target entity; According to the sub-event nodes associated with the candidate entity, one or more candidate sub-event nodes associated with the target sub-event node and the event association relationship between the target sub-event node and the candidate sub-event nodes are determined.

4. The knowledge system model for ship behavior analysis according to claim 1, It is characterized in that The theme event layer further includes: a theme event determination module, which is used to determine the theme event corresponding to each event sub-node according to the following steps: Through deep learning and information extraction algorithms, event semantic information is extracted from the sub-event nodes of the target news report; Determining the subject event to which the target news report sub-event node belongs according to the event semantic information; According to the event association relationship of the target news report sub-event node, the subject event to which the target news report sub-event node belongs is determined as the subject event of the sub-event node associated with the target news report sub-event node.

5. The knowledge system model for ship behavior analysis according to claim 1, It is characterized in that The ship behavior analysis model includes: a ship behavior dynamic identification model, a ship movement prediction model, and a ship behavior abnormality warning model; The knowledge system model also includes: A training module is used to train the ship behavior analysis model based on a machine learning algorithm using the real-time dynamic data stored in the subject event layer.

6. The knowledge system model for ship behavior analysis according to claim 1, It is characterized in that The ship behavior analysis model is stored in the active model layer as an independent entity; The model attribute information of the ship behavior analysis model includes: model unique identification ID, model function name, model type, model file, input parameters, output parameters, creation time and model function body description.

7. The knowledge system model for ship behavior analysis according to claim 1, It is characterized in that The knowledge system model also includes: A subject event layer updating module, used for updating the sub-event nodes and event association relationships in the real-time dynamic data at intervals of a first time length; The regularity knowledge updating module is used to update the regularity knowledge of the ship activities according to the real-time dynamic data.

8. A ship behavior analysis method, It is characterized in that The knowledge system model for ship behavior analysis applied to any one of claims 1 to 7 comprises: Receive target subject event query information input by the owner user; Input the target subject event query information into the ship behavior analysis model, use the ship activity law class knowledge, and generate the target ship behavior analysis result during the occurrence of the target subject event according to the real-time dynamic data related to the target subject event extracted from the subject event layer; The behavior analysis result of the target ship is displayed to the business user, or the sub-events in various fields related to the target theme event are connected in series according to the time element.

9. An electronic device, It is characterized in that It comprises a memory, a processor and a computer program stored in the memory, and the processor executes the computer program to implement the ship behavior analysis method described in claim 8.

10. A computer-readable storage medium, It is characterized in that A computer program / instruction is stored thereon, and when the computer program / instruction is executed by a processor, the ship behavior analysis method described in claim 8 is implemented.

Citation Information

Patent Citations

  • Ship industry knowledge graph construction and analysis method

    CN113987210A

  • Knowledge reasoning-oriented ship behavior formalized expression method and system

    CN114443978A

  • Ship activity knowledge graph construction method

    CN114647736A