Shipping risk assessment method

Abstract

The invention relates to a shipping risk assessment method, which comprises the following steps of: S1, establishing a semantic network of entities related to shipping and relationships among the entities, and determining a shipping basic information knowledge graph; s2, acquiring multi-source data related to shipping risks, and extracting risk entities and risk information from the multi-source data; s3, expanding a shipping basic information knowledge graph according to the risk entity and the risk information, constructing a risk association knowledge graph, and obtaining relationships between different risk factors and ship risks; s4, the risk factors are clustered, a risk identification model is constructed according to a clustering result to carry out quantitative evaluation on risks, a shipping risk level is obtained according to a shipping risk inference rule, if the shipping risk level is a low risk or a high risk, direct output is carried out, and otherwise, S5 is carried out; and S5, based on a quantitative evaluation result, carrying out weighted average calculation on the multi-factor risk to obtain a risk index, and determining a shipping risk level of the ship according to the risk index. Compared with the prior art, the method has the advantages of being high in precision and reliability.

Description

Technical Field

[0001] The present invention relates to the field of shipping risk assessment, and in particular to a shipping risk assessment method. Background Art

[0002] With the rapid development and widespread application of big data technology and artificial intelligence, the shipping industry is entering a new era of intelligent transformation. In the complex environment of global trade, shipping companies face a variety of risks, including meteorological disasters, navigation safety, market fluctuations, etc. In order to improve the efficiency of risk management, data-driven risk identification and assessment methods have gradually attracted attention. As an advanced technical means, knowledge graphs can integrate, analyze and reason a large amount of heterogeneous data in a structured and semantic way, providing intelligent risk identification and prediction support for the shipping industry.

[0003] In the shipping industry, risk factors are extremely complex and interrelated. Traditional risk management methods usually rely on experience and predefined rules, which are difficult to handle large-scale and dynamically changing risk information. This has led to increasingly higher requirements for the accuracy and real-time performance of risk identification and assessment. By building a system based on knowledge graphs, risk data from different sources can be effectively integrated, potential risk relationships can be identified, and complex risk networks can be made clearer, providing strong support for decision-making.

[0004] In recent years, shipping risk identification and assessment methods are mainly carried out through data statistics and model analysis, usually relying on manually set rules and algorithms. These traditional methods usually require a large number of pre-processing steps, including data cleaning, feature extraction, risk classification, etc., and often show certain limitations when dealing with changing shipping environments (such as weather changes, market fluctuations, operational errors, etc.). Common risk identification methods include tools such as Bayesian networks and risk scoring cards, but they are less versatile and difficult to adapt to complex scenarios and changing risk factors.

[0005] Since traditional methods mostly rely on empirical rules and are limited by subjective judgment and specific situations, they are not effective in dealing with different types of risks (such as ship accidents, port delays, and natural disasters). Especially in the face of complex shipping environments, traditional methods are often unable to quickly and accurately identify potential risks. In addition, the diversity and dynamic changes in shipping data (such as real-time meteorological data, ship trajectory data, and market fluctuation information) make it difficult to deal with risk identification based on fixed rules, resulting in low accuracy and timeliness of risk assessment. Summary of the invention

[0006] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a shipping risk assessment method with high accuracy and reliability.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] A shipping risk assessment method, comprising:

[0009] S1. Establish a semantic network of entities related to shipping and the relationships between entities, and determine the knowledge graph of basic shipping information;

[0010] S2. Obtain multi-source data related to shipping risks and extract risk entities and risk information from the multi-source data;

[0011] S3. Expand the basic shipping information knowledge graph according to the extracted risk entities and risk information, construct a risk association knowledge graph, and obtain the relationship between different risk factors and ship risks;

[0012] S4. Cluster the risk factors using a clustering algorithm, and build a risk identification model based on the clustering results to quantitatively assess the risks. The shipping risk level is obtained according to the shipping risk reasoning rules. If the shipping risk level is low risk or high risk, it is directly output, otherwise it goes to S5;

[0013] S5. Based on the quantitative assessment results of S4, the risk index is calculated by weighted average of multiple risk factors, and the shipping risk level of the ship is determined according to the risk index.

[0014] Preferably, the semantic network of shipping-related entities and relationships between entities is established in S1 to determine the knowledge graph of basic shipping information, specifically including:

[0015] Define shipping related entities E = {E1, E2, ..., E n}, the entities include ship entities, port entities, route entities, meteorological entities and crew entities;

[0016] Determine the entity E i Properties in, For entity E i The mth i Attributes;

[0017] Construct a mapping relationship network between entities. For entity E i With entity E j , the mapping relationship is expressed as R(E i , E j )={R1,R2,...,R p}, where R p For entity E i With entity E j The pth mapping result between ;

[0018] Use the resource description framework to describe entities and their relationships, where entity E i With entity E j The resource description framework relationship network structure between them is represented as RDFs=(E i , R, E j );

[0019] Constructing the basic shipping information knowledge graph G = (E, A E , A G , W E , R E ), where E is the entity node set, A E ={A(E1), A(E2), ..., A(E n )} is the attribute set of all entity nodes, A G =[a ij ], (e i , e j )∈E indicates that entity node i is associated with entity node j, W E Indicates the strength of the entity relationship, R E is the relationship type matrix.

[0020] Preferably, the multi-source data related to shipping risks include meteorological data, shipping accident reports, ship information and route information, wherein the shipping accident reports contain human factors.

[0021] Preferably, the step S2 of acquiring multi-source data related to shipping risks and extracting risk entities and risk information from the multi-source data specifically includes:

[0022] S21. Match multi-source data according to timestamp attributes and geographic coordinate attributes, and convert multi-source data into a unified representation Among them, U k is the unified representation of the kth data, t is the timestamp, x and y are the geographic coordinates, The kth data is unified into the Mth k Attributes, M k is the number of attributes in the k-th data;

[0023] S22. Use the trained conditional random field model to identify risk entities and risk information in shipping accident reports.

[0024] Preferably, the trained conditional random field model is used in S22 to identify risk entities and risk information in the shipping accident report. The pre-training process of the conditional random field model is specifically as follows:

[0025] Use named entity recognition technology to label each word in the shipping accident report, where the label is the set entity category;

[0026] The obtained text data T is converted into a feature vector F = {f1, f2, ..., f s}, where f s represents the sth feature;

[0027] The labeled data set is used to train the conditional random field model, and the mathematical expression of the conditional random field model is:

[0028]

[0029] Among them, λ k is the weight of the kth feature, f k is the kth feature, u is the number of words in the input sequence x, and Z(x) is the normalization factor.

[0030] Preferably, in S3, the basic shipping information knowledge graph is expanded according to the extracted risk entities and risk information, a risk association knowledge graph is constructed, and the relationship between different risk factors and ship risks is obtained, including the following sub-steps:

[0031] S31, expanding the basic shipping information knowledge graph according to the extracted risk entities and risk information, and constructing a risk association knowledge graph;

[0032] S32, extract the risk factors that affect the safety of ship navigation under different weather conditions, sea conditions and crew conditions, and establish a random variable set {X1, X2, ..., X q}, where X q Corresponding to the qth shipping risk factor, the shipping risk factors include wind speed, wave height, ship type, channel conditions, sea depth, crew experience, crew working hours, crew fatigue and operating error rate;

[0033] S33. Use the Bayesian network to obtain the impact of different weather factors, sea conditions and human factors on navigation safety, and select the risk level with the maximum posterior probability as the final predicted risk level;

[0034] S34. Use the Bayesian network risk assessment results obtained in S33 as input for querying and reasoning about the relationship between different factors under complex weather and sea conditions, and use SPARQL to query the relationship between multiple factors and ship risks under different weather and sea conditions, wherein the multiple factors include ship age, ship type and channel congestion.

[0035] Preferably, in S4, a clustering algorithm is used to cluster the risk factors, specifically the following sub-steps:

[0036] S401, using one-hot encoding to extract features and convert different shipping risk factors into feature vectors;

[0037] S402, cluster the shipping risk factors using an incomplete multi-view clustering algorithm based on self-representation, capture the potential correlation between data from different perspectives, and output the clustered risk feature vector C = {C1, C2, ..., C r}, where C r is the rth feature group after clustering, indicating the feature cluster that is closely related to risk under specific conditions;

[0038] The incomplete multi-view clustering algorithm based on self-representation utilizes the subspace information shared by multiple views, learns the consistency matrix S to save the effective consistent information of multiple views, and assigns different weights w to the views. v , the objective function expression is:

[0039]

[0040]

[0041] Among them, X v is the matrix of the vth view, d v is the feature dimension of the vth perspective, N is the number of samples, represents the known data matrix of the vth viewing angle, is the extraction matrix of the vth view, H v is the representation matrix of the vth view.

[0042] Preferably, the risk identification model is constructed according to the clustering results to quantitatively evaluate the risk, and the shipping risk level is obtained according to the shipping risk reasoning rules, which specifically includes the following sub-steps:

[0043] S411, construct a risk identification model, use the risk feature vector C and consistency matrix S obtained by clustering as input data, and output the risk quantitative assessment result;

[0044] S412, constructing a shipping risk reasoning rule by using knowledge reasoning to reason about shipping risks, wherein the shipping risk reasoning rule includes a weather condition reasoning sub-rule, a ship status reasoning sub-rule, an environmental condition reasoning sub-rule and a human factor reasoning sub-rule;

[0045] S413. According to the correlation between shipping risk factors, the judgment results of each inference sub-rule are comprehensively inferred and judged to obtain the inference result of shipping risk level with multiple factors superimposed. If the inference results of each inference sub-rule are all low risk, the shipping risk level is judged to be low risk and directly output. If the inference result of any inference sub-rule is high risk, the shipping risk level is judged to be high risk and directly output. In other cases, go to S5.

[0046] Preferably, the shipping risk reasoning rules in S412 specifically include:

[0047] Weather condition reasoning sub-rules, including:

[0048] If the wind speed is > A1km / h and the wave height is > B meters and the route is in open waters, it is judged to be high risk;

[0049] If the wind speed is ≤A1km / h and the wave height is ≤B meters and the route is a coastal route, it is judged to be medium risk; and

[0050] If the wind speed is ≤A2km / h and the weather is clear and the route is an inland river route, it is judged to be low risk;

[0051] Ship status reasoning sub-rules, including:

[0052] If the ship type is a small ship and the load is > C% and the wind speed is > A3km / h, it is judged as high risk;

[0053] If the vessel type is a large cargo ship and it is sailing in a safe channel and the wind speed is ≤A3km / h, it is judged to be low risk; and

[0054] If the ship type is a large cargo ship and the wind speed is > A1km / h and it is sailing in open waters, it is judged to be medium risk;

[0055] Environmental condition reasoning sub-rules, including:

[0056] If the sea area is shallow and the current speed is > Dm / s and the visibility is < E1km, it is judged as high risk;

[0057] If the route is a safe channel and visibility> E2km, it is judged as low risk; and

[0058] If the route is a busy waterway and visibility is less than E3km, it is considered to be medium risk;

[0059] And the human factor reasoning sub-rules include:

[0060] If the crew working hours are greater than F1 hours and the crew fatigue is high and the ship type is a small ship, it is judged as high risk;

[0061] If the crew experience is > G years and the crew fatigue is low and the operational error rate is < H1%, it is judged as low risk; and

[0062] If the crew working hours are ≤F2 hours and the operational error rate is >H2% and the weather conditions are high risk, it is judged as medium risk;

[0063] Among them, A1, A2, A3, B, C, D, E1, E2, E3, F1, F2, G, H1 and H2 are set constants, and satisfy A1>A3>A2, E2>E1>E3, F1>F2, H2>H1.

[0064] Preferably, in S5, based on the quantitative assessment result of S4, a risk index is calculated by weighted average of the multi-factor risks, and the navigation risk level of the ship is determined according to the risk index. The calculation expression of the risk index α is:

[0065]

[0066] Among them, α i is the corresponding risk quantification assessment result of the i-th shipping risk factor, quantifying high risk as 9, medium risk as 5, and low risk as 2, w i is the weight of the ith shipping risk factor, obtained using the regression model, and

[0067] For the risk index α, if α∈[0,3), it is judged as low risk, if α∈[3,7), it is judged as medium risk, and if α∈[7,10], it is high risk.

[0068] Compared with the prior art, the present invention has the following beneficial effects:

[0069] (1) The present invention defines the attributes and relationships of shipping-related entities, constructs a knowledge graph, and integrates shipping multi-source data including meteorological data, shipping accident reports, ship information, route information, and human factors. This realizes the effective association between shipping multi-source data, can comprehensively reflect various factors affecting the safety of ship navigation, greatly improves the understanding and assessment capabilities of shipping risks, and ensures more scientific decision-making.

[0070] (2) The present invention constructs a risk identification model based on the risk factor clustering results, uses shipping risk inference rules to quantitatively evaluate the shipping risk level, and combines knowledge graphs with machine learning to deeply capture the complex relationship between shipping risk factors. For example, it associates weather conditions, navigation environment, crew conditions and historical accident data, provides contextual support, and significantly improves the accuracy of shipping risk prediction.

[0071] (3) The present invention uses a reasoning mechanism based on knowledge graphs to perform complex risk reasoning operations through established entity relationships, and automatically infers risk associations under specific conditions. For example, the risk exposure levels of different types of ships under specific weather conditions provide a practical method for discovering potential risks, and add an important basis for the scientific and systematic nature of the shipping decision-making process, ensuring that shipping management is more efficient and safe.

[0072] (4) An incomplete multi-view clustering algorithm based on self-representation is used to cluster shipping risk factors, capturing the potential correlation between data from different perspectives, improving the mining and utilization of existing shipping data information, and further improving the accuracy of subsequent shipping assessment results. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 is a flow chart of the method of the present invention;

[0074] Figure 2 A schematic diagram of the shipping risk prediction and quantitative assessment process;

[0075] Figure 3 Schematic diagram of the risk assessment process. DETAILED DESCRIPTION

[0076] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0077] Example

[0078] like Figure 1 As shown, this embodiment provides a shipping risk assessment method, which can effectively optimize the bottleneck problem of shipping risk information dispersion and risk identification and assessment accuracy. The method includes the following steps:

[0079] S1. Establish a semantic network of entities related to shipping and the relationships between entities, and determine the knowledge graph of basic shipping information;

[0080] S2. Obtain multi-source data related to shipping risks and extract risk entities and risk information from the multi-source data;

[0081] S3. Expand the basic shipping information knowledge graph according to the extracted risk entities and risk information, construct a risk association knowledge graph, and obtain the relationship between different risk factors and ship risks;

[0082] S4. Cluster the risk factors using a clustering algorithm, and build a risk identification model based on the clustering results to quantitatively assess the risks. The shipping risk level is obtained according to the shipping risk reasoning rules. If the shipping risk level is low risk or high risk, it is directly output, otherwise it goes to S5;

[0083] S5. Based on the quantitative assessment results of S4, the risk index is calculated by weighted average of multiple risk factors, and the shipping risk level of the ship is determined according to the risk index.

[0084] Next, the method of this embodiment is introduced in detail.

[0085] S1. Establish a semantic network of entities related to shipping and the relationships between entities, and determine the knowledge graph of basic shipping information, including:

[0086] S11. Define shipping related entities E = {E1, E2, ..., E n}, entities include ship entity, port entity, route entity, weather entity and crew entity; for example, E1=ship, E2=port, E3=route, E4=weather, E5=crew, etc.

[0087] S12. Determine entity E i Properties in, For entity E i The mth i Attributes;

[0088] For example, A(ship) = {load, length, ...}, A(port) = {throughput, geographical location, ...}, A(crew) = {crew experience, crew working hours, ...}, etc.

[0089] S13, build a mapping relationship network between entities, for entity E i With entity E j , the mapping relationship is expressed as R(E i , E j )={R1,R2,...,R p}, where R p For entity E i With entity E j The p-th mapping result between .

[0090] For "ship-docking-port", "ship-sailing-route", "route-affected-weather", etc., R(ship, port) = docking, R(ship, route) = sailing, R(route, weather) = affected, etc.

[0091] S14. Use the resource description framework RDF to describe entities and relationships between entities, where entity Ei With entity E j The resource description framework relationship network structure between them is represented as RDFs=(E i , R, E j ).

[0092] S15. Constructing the basic shipping information knowledge graph G = (E, A E , A G , W E , R E ), where E is the entity node set, A E ={A(E1), A(E2), ..., A(E n )} is the attribute set of all entity nodes, A G =|a ij ], (e i , e j )∈E indicates that entity node i is associated with entity node j, W E Indicates the strength of the entity relationship, R E is the relationship type matrix.

[0093] S2. Obtain multi-source data related to shipping risks, extract risk entities and risk information from the multi-source data, and perform pre-processing. The multi-source data related to shipping risks include meteorological data, shipping accident reports, ship information and route information, etc. Shipping accident reports contain human factors. The specific sub-steps include the following:

[0094] Before processing multi-source data, missing value filling and outlier processing are performed to unify the format.

[0095] S21. Match multi-source data according to timestamp attributes and geographic coordinate attributes, and convert multi-source data into a unified representation Among them, U k is the unified representation of the kth data, t is the timestamp, x and y are the geographic coordinates, The kth data is unified into the Mth k Attributes, M k is the number of attributes in the k-th data;

[0096] S22. Use command entity recognition technology to extract shipping risk information from shipping accident reports, and use the trained conditional random field model to identify important risk entities and risk information such as ship names, ports, weather conditions, etc. in shipping accident reports. For example, identify important entities such as ship names, ports, weather conditions, etc. in shipping reports.

[0097] Among them, the pre-training process of the conditional random field model is specifically as follows:

[0098] Named entity recognition (NER) is used to label each word in the shipping accident report, where the label is the set entity category (such as "ship status", "port conditions", "weather conditions" or "human factors");

[0099] The obtained text data T is converted into a feature vector F = {f1, f2, ..., f s}, where f s represents the sth feature, which usually includes the word itself, part of speech, prefix and suffix of the word, whether it is capitalized, digital features, context words, etc.;

[0100] The obtained text data T is converted into a feature vector F = {f1, f2, ..., f s}, where f s represents the sth feature;

[0101] The labeled data set is used to train the conditional random field model, and the mathematical expression of the conditional random field model is:

[0102]

[0103] Among them, λ k is the weight of the kth feature, f k is the kth feature, u is the number of words in the input sequence x, and Z(x) is a normalization factor that ensures that the sum of the probabilities of all possible label sequences is 1.

[0104] S3. Expand the basic shipping information knowledge graph based on the extracted risk entities and risk information, construct a risk association knowledge graph, and obtain the relationship between different risk factors and ship risks, such as Figure 2 As shown, it specifically includes the following sub-steps:

[0105] S31. Expand the basic shipping information knowledge graph based on the extracted risk entities and risk information, and construct the risk association knowledge graph KG2, KG2 = KG1∪E ner ∪R risk , KG1 is the basic information knowledge graph of shipping, E ner is the set of risk entities identified from shipping accident reports, R risk is a set of relationships between risk entities identified from shipping accident reports;

[0106] S32, extract the risk factors that affect the safety of ship navigation under different weather conditions, sea conditions and crew conditions, and establish a random variable set {X1, X2, ..., X q}, where X qCorresponding to the qth shipping risk factor, the shipping risk factors include wind speed, wave height, ship type, channel conditions, sea depth, crew experience, crew working hours, crew fatigue and operating error rate, for example:

[0107] X1 = wind speed (e.g. {low, medium, high})

[0108] X2 = wave height (e.g. {low, medium, high})

[0109] X3 = ship type (e.g. {small ship, medium ship, large ship})

[0110] X4 = Crew experience (eg {junior, intermediate, senior})

[0111] X5 = crew working hours (e.g. {short, moderate, long})

[0112] X6 = crew fatigue (e.g. {low, medium, high})

[0113] …

[0114] X q =Other risk factors (such as channel conditions, sea depth, etc.)

[0115] S33. Use Bayesian network to obtain the impact of different weather factors, sea conditions and human factors on navigation safety, and select the risk level with the maximum posterior probability risk * As the final predicted risk level, the calculation expression is:

[0116]

[0117] Among them, P(risk) is the prior probability, P(X i |risk) is the conditional probability of each risk factor given the risk level risk;

[0118] S34. Use the Bayesian network risk assessment results obtained in S33 as input for querying and reasoning about the relationship between different factors under complex weather and sea conditions, and use SPARQL to query the relationship between multiple factors (such as ship age, ship type, and channel congestion) and ship risks under different weather and sea conditions.

[0119] S4: Cluster the risk factors using clustering algorithm, and build a risk identification model based on the clustering results to quantitatively assess the risks. The shipping risk level is obtained according to the shipping risk reasoning rules. If the shipping risk level is low risk or high risk, it is directly output and a risk assessment report is formed. Otherwise, go to S5, including:

[0120] Clustering algorithm is used to cluster risk factors, with the following sub-steps:

[0121] S401, using one-hot encoding to extract features, converting different shipping risk factors (such as weather conditions, ship types, sailing time, crew working hours, etc.) into feature vectors X = [x1, x2, ..., x n ];

[0122] S402, cluster similar shipping risk factors using an incomplete multi-view clustering algorithm based on self-representation, capture the potential correlation between data from different perspectives, and output the clustered risk feature vector C = {C1, C2, ..., C r}, where C r is the rth feature group after clustering, indicating the feature cluster that is closely related to risk under specific conditions.

[0123] Due to the complex associations between shipping risk factors, such as different weather conditions may affect different types of ships, a self-representation model is added to the spectral clustering algorithm in the basic missing case. The model represents or reconstructs data using its own linear combination, which can effectively capture the potential correlation between data from different perspectives. For example, certain types of ships may be more vulnerable to risks under different sea areas and weather conditions.

[0124] In this embodiment, the incomplete multi-view clustering algorithm IMCS based on self-representation uses the subspace information shared by multiple views to learn the consistency matrix S to save the effective consistent information of multiple views and assign different weights w to the views. v , the objective function expression is:

[0125]

[0126]

[0127] Among them, X v is the matrix of the vth view, d v is the feature dimension of the vth perspective, N is the number of samples, represents the known data matrix of the vth viewing angle, is the extraction matrix of the vth view, H v is the representation matrix of the vth view.

[0128] Based on the clustering results, a risk identification model is constructed to quantitatively evaluate the risks, and the shipping risk level is obtained according to the shipping risk reasoning rules, which specifically includes the following sub-steps:

[0129] S411, construct a risk identification model, use the risk feature vector C and consistency matrix S obtained by clustering as input data, and output the risk quantitative assessment result;

[0130] The risk identification model is trained using the feature vectors and risk labels (low risk, medium risk, high risk) in the clustering results as training data.

[0131] S412. Use knowledge reasoning technology to construct shipping risk reasoning rules based on factors such as weather conditions, navigation environment and crew conditions to reason about shipping risks. The shipping risk reasoning rules include weather condition reasoning sub-rules, ship status reasoning sub-rules, environmental condition reasoning sub-rules and human factor reasoning sub-rules;

[0132] 1) Weather condition reasoning sub-rules, such as wind speed, wave height, visibility and other meteorological conditions, are comprehensively evaluated according to different conditions and route types, including:

[0133] If the wind speed is > A1 km / h (in this embodiment, A1 = 50) and the wave height is > B meters (in this embodiment, b = 5) and the route is in open sea, it is determined to be high risk;

[0134] If the wind speed is ≤A1 km / h (in this embodiment, A1=50) and the wave height is ≤B meters (in this embodiment, b=5) and the route is a coastal route, it is determined to be a medium risk; and

[0135] If the wind speed is ≤ A2 km / h (in this embodiment, A2 = 20) and the weather is clear and the route is an inland river route, it is judged to be low risk;

[0136] 2) Ship status reasoning sub-rules, such as ship load, ship type, sailing speed and other factors, combined with weather conditions and sailing environment for comprehensive evaluation, including:

[0137] If the ship type is a small ship and the load is greater than C% (in this embodiment, C=90) and the wind speed is greater than A3 km / h (in this embodiment, A3=30), it is determined to be a high risk;

[0138] If the ship type is a large cargo ship and is sailing in a safe channel and the wind speed is ≤ A3 km / h (in this embodiment, A3 = 30), it is judged as low risk; and

[0139] If the ship type is a large cargo ship and the wind speed is greater than A1 km / h (in this embodiment, A1 = 50) and it is sailing in open waters, it is determined to be a medium risk;

[0140] 3) Environmental condition reasoning sub-rules, such as sea area characteristics, waterway conditions, etc., including:

[0141] If the sea area is shallow water and the sea current speed is greater than D m / s (in this embodiment, D=3) and the visibility is less than E1 km (in this embodiment, E1=2), it is determined to be high risk;

[0142] If the route is a safe channel and the visibility is greater than E2 km (in this embodiment, E2=5), it is determined to be a low risk; and

[0143] If the route is a busy waterway and visibility is less than E3 km (in this embodiment, E3 = 1), it is determined to be a medium risk;

[0144] 4) Human factor reasoning sub-rules, such as crew working hours, experience, fatigue, operating error rate and other factors, combined with ship status and weather conditions for evaluation include:

[0145] If the crew working hours (daily working hours) is greater than F1 hours (in this embodiment, F1=12) and the crew fatigue is high and the ship type is a small ship, it is determined to be a high risk;

[0146] If the crew experience is greater than G years (in this embodiment, G=10), the crew fatigue is low, and the operation error rate is less than H1% (in this embodiment, H1=5), it is determined to be low risk; and

[0147] If the crew working hours are ≤ F2 hours (in this embodiment, F2=8) and the operating error rate is > H2% (in this embodiment, H2=10) and the weather conditions are high risk, it is determined to be medium risk.

[0148] S413. According to the correlation between shipping risk factors, the judgment results of each inference sub-rule are comprehensively inferred and judged to obtain the inference result of shipping risk level with multiple factors superimposed, such as Figure 3 As shown, if the inference results of each inference sub-rule are low risk, the shipping risk level is determined to be low risk and directly output; if the inference result of any inference sub-rule is high risk, the shipping risk level is determined to be high risk and directly output; otherwise, go to S5.

[0149] S5. Based on the quantitative assessment results of S4, the risk index is calculated by weighted average of multiple risk factors, and the shipping risk level of the ship is determined according to the risk index.

[0150] Based on the quantitative assessment results of S4, the risk index is calculated by weighted average of multi-factor risks. The navigation risk level of the ship is determined according to the risk index. The calculation expression of the risk index α is:

[0151]

[0152] Among them, α i is the corresponding risk quantification assessment result of the i-th shipping risk factor, quantifying high risk as 9, medium risk as 5, and low risk as 2, w i is the weight of the ith shipping risk factor, obtained using the regression model, and

[0153] For the risk index α, if α∈[0,3), it is judged as low risk, if α∈[3,7), it is judged as medium risk, and if α∈[7,10], it is high risk.

[0154] In this embodiment, the risk index is recalculated at regular intervals, the risk level is updated according to the new risk index, and the risk status is continuously monitored.

[0155] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A shipping risk assessment method, characterized in that: include: S1. Establish a semantic network of entities related to shipping and the relationships between entities, and determine the knowledge graph of basic shipping information; S2. Obtain multi-source data related to shipping risks and extract risk entities and risk information from the multi-source data; S3. Expand the basic shipping information knowledge graph according to the extracted risk entities and risk information, construct a risk association knowledge graph, and obtain the relationship between different risk factors and ship risks; S4. Cluster the risk factors using a clustering algorithm, and build a risk identification model based on the clustering results to quantitatively assess the risks. The shipping risk level is obtained according to the shipping risk reasoning rules. If the shipping risk level is low risk or high risk, it is directly output, otherwise it goes to S5; S5. Based on the quantitative assessment results of S4, the risk index is calculated by weighted average of multiple risk factors, and the shipping risk level of the ship is determined according to the risk index.

2. A shipping risk identification and assessment method according to claim 1, characterized in that: In S1, a semantic network of shipping-related entities and relationships between entities is established to determine the knowledge graph of basic shipping information, including: Define shipping related entities E = {E1, E2, ..., E n }, the entities include ship entities, port entities, route entities, meteorological entities and crew entities; Determine the entity E i Properties in, For entity E i The mth i Attributes; Construct a mapping relationship network between entities. For entity E i With entity E j , the mapping relationship is expressed as R(E i ,E j )={R1,R2,…,R p }, where R p For entity E i With entity E j The pth mapping result between ; Use the resource description framework to describe entities and their relationships, where entity E i With entity E j The resource description framework relationship network structure between them is represented as RDFs=(E i ,R,E j ); Constructing the basic shipping information knowledge graph G = (E, A E ,A G ,W E ,R E ), where E is the entity node set, A e ={A(E1),A(E2),…,A(E n )} is the attribute set of all entity nodes, A G =[a ih ], (e i ,e j )∈E indicates that entity node i is associated with entity node j, W E Indicates the strength of the entity relationship, R E is the relationship type matrix.

3. A shipping risk identification and assessment method according to claim 1, characterized in that: The multi-source data related to shipping risks include meteorological data, shipping accident reports, ship information and route information, wherein the shipping accident reports contain human factors.

4. A shipping risk identification and assessment method according to claim 3, characterized in that: The S2 acquires multi-source data related to shipping risks and extracts risk entities and risk information from the multi-source data, specifically including: S21. Match multi-source data according to timestamp attributes and geographic coordinate attributes, and convert multi-source data into a unified representation Among them, U k is the unified representation of the kth data, t is the timestamp, x and y are the geographic coordinates, The kth data is unified into the Mth data. k Attributes, M k is the number of attributes in the k-th data; S22. Use the trained conditional random field model to identify risk entities and risk information in shipping accident reports.

5. A shipping risk identification and assessment method according to claim 3, characterized in that: In S22, the trained conditional random field model is used to identify risk entities and risk information in the shipping accident report. The pre-training process of the conditional random field model is specifically as follows: Use named entity recognition technology to label each word in the shipping accident report, where the label is the set entity category; The obtained text data T is converted into a feature vector F = {f1, f2, ..., f s }, where f s represents the sth feature; The labeled data set is used to train the conditional random field model, and the mathematical expression of the conditional random field model is: Among them, λ k is the weight of the kth feature, f k is the kth feature, u is the number of words in the input sequence x, and Z(x) is the normalization factor.

6. A shipping risk identification and assessment method according to claim 3, characterized in that: In S3, the basic shipping information knowledge graph is expanded according to the extracted risk entities and risk information, a risk association knowledge graph is constructed, and the relationship between different risk factors and ship risks is obtained, including the following sub-steps: S31, expanding the basic shipping information knowledge graph according to the extracted risk entities and risk information, and constructing a risk association knowledge graph; S32, extract the risk factors that affect the safety of ship navigation under different weather conditions, sea conditions and crew conditions, and establish a random variable set {X1, X2, …, X q }, where X q corresponding to the qth shipping risk factor, the shipping risk factors include wind speed, wave height, ship type, channel conditions, sea depth, crew experience, crew working hours, crew fatigue and operating error rate; S33. Use the Bayesian network to obtain the impact of different weather factors, sea conditions and human factors on navigation safety, and select the risk level with the maximum posterior probability as the final predicted risk level; S34. Use the Bayesian network risk assessment results obtained in S33 as input for querying and reasoning about the relationship between different factors under complex weather and sea conditions, and use SPARQL to query the relationship between multiple factors and ship risks under different weather and sea conditions, wherein the multiple factors include ship age, ship type and channel congestion.

7. A shipping risk identification and assessment method according to claim 1, characterized in that: In S4, a clustering algorithm is used to cluster risk factors, specifically the following sub-steps: S401, using one-hot encoding to extract features and convert different shipping risk factors into feature vectors; S402, cluster the shipping risk factors using an incomplete multi-view clustering algorithm based on self-representation, capture the potential correlation between data from different perspectives, and output the risk feature vector C after clustering = {C1, C2, ..., C r }, where C r is the rth feature group after clustering, indicating the feature cluster that is closely related to risk under specific conditions; The incomplete multi-view clustering algorithm based on self-representation utilizes the subspace information shared by multiple views, learns the consistency matrix S to save the effective consistent information of multiple views, and assigns different weights w to the views. v , the objective function expression is: Among them, X v is the matrix of the vth view, d v is the feature dimension of the vth perspective, N is the number of samples, X E v represents the known data matrix of the vth viewing angle, is the extraction matrix of the vth view, H v is the representation matrix of the vth view.

8. A shipping risk identification and assessment method according to claim 7, characterized in that: In S4, a risk identification model is constructed based on the clustering results to quantitatively evaluate the risks, and the shipping risk level is obtained according to the shipping risk reasoning rules, which specifically includes the following sub-steps: S411. Construct a risk identification model, take the risk feature vector C and the consistency matrix S obtained by clustering as input data, and output the risk quantitative evaluation result; S412. Use knowledge reasoning to construct shipping risk reasoning rules to reason about shipping risks. The shipping risk reasoning rules include weather condition reasoning sub-rules, ship status reasoning sub-rules, environmental condition reasoning sub-rules, and human factor reasoning sub-rules; S413. According to the relevance between shipping risk factors, comprehensively reason and judge the determination results of each reasoning sub-rule to obtain the shipping risk level reasoning result of multi-factor superposition. If the reasoning results of each reasoning sub-rule are all low risks, it is determined that the shipping risk level is low risk and directly output. If the reasoning result of any reasoning sub-rule is high risk, it is determined that the shipping risk level is high risk and directly output. In other cases, go to S5.

9. A shipping risk identification and assessment method according to claim 8, characterized in that: The shipping risk reasoning rules in S412 specifically include: Weather condition reasoning sub-rules, including: If the wind speed > A1 km / h and the wave height > B meters and the route is an open sea area, it is determined to be a high risk; If the wind speed ≤ A1 km / h and the wave height ≤ B meters and the route is a coastal route, it is determined to be a medium risk; and If the wind speed ≤ A2 km / h and the weather is clear and the route is an inland river route, it is determined to be a low risk; Ship status reasoning sub-rules, including: If the ship type is a small ship and the load > C% and the wind speed > A3 km / h, it is determined to be a high risk; If the ship type is a large cargo ship and it is sailing in a safe channel and the wind speed ≤ A3 km / h, it is determined to be a low risk; and If the ship type is a large cargo ship and the wind speed > A1 km / h and it is sailing in an open sea area, it is determined to be a medium risk; Environmental condition reasoning sub-rules, including: If the sea area is shallow water area and the sea current speed > D m / s and the visibility < E1 km, it is determined to be a high risk; If the route is a safe channel and the visibility > E2 km, it is determined to be a low risk; and If the route is a busy channel and the visibility < E3 km, it is determined to be a medium risk; And human factor reasoning sub-rules, including: If the crew working hours > F1 hours and the crew fatigue level is high and the ship type is a small ship, it is determined to be a high risk; If the crew experience > G years and the crew fatigue level is low and the operation error rate < H1%, it is determined to be a low risk; and If the crew working hours ≤ F2 hours and the operation error rate > H2% and the weather condition is high risk, it is determined to be a medium risk; Among them, A1, A2, A3, B, C, D, E1, E2, E3, F1, F2, G, H1, and H2 are set constants, and satisfy A1 > A3 > A2, E2 > E1 > E3, F1 > F2, H2 > H1.

10. A shipping risk identification and assessment method according to claim 8, characterized in that: In S5, based on the quantitative evaluation result of S4, the risk index is calculated by weighted average of multi-factor risks, and the navigation risk level of the ship is determined according to the risk index. The calculation expression of the risk index α is: Among them, α i is the corresponding risk quantification assessment result of the i-th shipping risk factor, quantifying high risk as 9, medium risk as 5, and low risk as 2, w i is the weight of the ith shipping risk factor, obtained using the regression model, and For the risk index α, if α∈[0,3), it is judged as low risk, if α∈[3,7), it is judged as medium risk, and if α∈[7,10], it is high risk.

Images