Sea ice risk prediction method and device, computer equipment, storage medium and product

Abstract

The invention discloses a sea ice risk prediction method and device, computer equipment, a storage medium and a product. The method comprises the following steps: acquiring historical environment data and historical sea ice monitoring data; constructing a linear regression relation model between the target environment factor and the sea ice data and a time sequence analysis model of the sea ice data according to the historical environment data and the historical sea ice monitoring data; fusing the linear regression relation model and the time sequence analysis model to obtain a hybrid model; and acquiring real-time environment data, and generating sea ice prediction data by using the hybrid model. The sea ice generation and change rule can be better described by combining the explanation capability of linear regression on influence factors and the dynamic grasp of a time sequence analysis model on historical data, the advantages of different models can be effectively integrated by introducing a hybrid model, the prediction precision of sea ice risks and the capture capability of trends are improved, and the prediction efficiency is improved. Therefore, more accurate sea ice risk prediction under a complex environment condition is realized.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of maritime safety technology, and in particular to a sea ice risk prediction method, device, computer equipment, storage medium and product. Background Art

[0002] Sea ice poses a serious threat to maritime shipping, marine engineering and other activities. Accurately predicting the formation and movement of sea ice is crucial to avoiding risks and formulating countermeasures. However, due to the complexity of sea ice formation, movement and melting, and its influence by various meteorological and oceanic conditions, the traditional single model prediction method has certain limitations in dealing with these complex dynamic processes, and it is difficult to achieve high-precision early warning effects. Summary of the invention

[0003] Embodiments of the present invention provide a sea ice risk prediction method, apparatus, computer equipment, storage medium and product to achieve more accurate sea ice risk prediction under complex environmental conditions.

[0004] In a first aspect, an embodiment of the present invention provides a method for predicting sea ice risk, the method comprising:

[0005] Obtain historical environmental data and historical sea ice monitoring data;

[0006] Constructing a linear regression relationship model between target environmental factors and sea ice data, and a time series analysis model for sea ice data based on the historical environmental data and the historical sea ice monitoring data;

[0007] The linear regression relationship model and the time series analysis model are integrated to obtain a hybrid model;

[0008] Real-time environmental data is acquired and sea ice prediction data is generated using the hybrid model.

[0009] Optionally, before constructing a linear regression relationship model between the target environmental factor and the sea ice data according to the historical environmental data and the historical sea ice monitoring data, the method further includes:

[0010] The target environmental factor having an influence on the sea ice data is selected from various environmental factors in the environmental data by using a Bayesian network or Granger causality analysis method.

[0011] Optionally, constructing a linear regression relationship model between the target environmental factor and the sea ice data, and a time series analysis model for the sea ice data according to the historical environmental data and the historical sea ice monitoring data includes:

[0012] Determining whether the historical sea ice monitoring data is a stationary time series;

[0013] If so, the time series analysis model is obtained based on the autoregressive moving average model;

[0014] If not, the historical sea ice monitoring data is differentially processed to be converted into a stationary time series, and the time series analysis model is obtained based on an autoregressive difference moving average model.

[0015] Optionally, the linear regression relationship model and the time series analysis model are fused to obtain a hybrid model, including:

[0016] A particle swarm optimization algorithm or a genetic algorithm is used to search for the optimal weight combination between the linear regression relationship model and the time series analysis model in the hybrid model.

[0017] Optionally, a penalty term of a preset physical constraint is set in the loss function of the hybrid model.

[0018] Optionally, after acquiring the real-time environmental data and generating sea ice prediction data using the hybrid model, the method further includes:

[0019] Acquire real-time sea ice monitoring data corresponding to the sea ice prediction data;

[0020] An online learning mechanism is adopted to update the hybrid model online in combination with the real-time sea ice monitoring data to redefine the target environmental factors and the weight combination between the linear regression relationship model and the time series analysis model.

[0021] In a second aspect, an embodiment of the present invention further provides a sea ice risk prediction device, the device comprising:

[0022] A data acquisition module, used to obtain historical environmental data and historical sea ice monitoring data;

[0023] A model building module, used to build a linear regression relationship model between the target environmental factor and the sea ice data, and a time series analysis model for the sea ice data based on the historical environmental data and the historical sea ice monitoring data;

[0024] A model fusion module, used to fuse the linear regression relationship model and the time series analysis model to obtain a hybrid model;

[0025] The real-time prediction module is used to obtain real-time environmental data and generate sea ice prediction data using the hybrid model.

[0026] In a third aspect, an embodiment of the present invention further provides a computer device, the computer device comprising:

[0027] one or more processors;

[0028] A memory for storing one or more programs;

[0029] When the one or more programs are executed by the one or more processors, the one or more processors implement the sea ice risk prediction method provided by any embodiment of the present invention.

[0030] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the sea ice risk prediction method provided by any embodiment of the present invention.

[0031] In a fifth aspect, an embodiment of the present invention further provides a computer program product, which includes a computer program, and when the program is executed by a processor, the sea ice risk prediction method provided by any embodiment of the present invention is implemented.

[0032] The embodiment of the present invention provides a sea ice risk prediction method, which first obtains historical environmental data and historical sea ice monitoring data, then constructs a linear regression relationship model between the target environmental factor and the sea ice data based on the obtained historical environmental data and historical sea ice monitoring data, and a time series analysis model of the sea ice data, and then fuses the linear regression relationship model and the time series analysis model to obtain a hybrid model, and then obtains real-time environmental data, and uses the hybrid model to generate sea ice prediction data. The sea ice risk prediction method provided by the embodiment of the present invention can better describe the laws of sea ice formation and change by combining the ability of linear regression to explain influencing factors and the dynamic grasp of historical data by the time series analysis model. By introducing a hybrid model, the advantages of different models can be effectively integrated to improve the prediction accuracy of sea ice risk and the ability to capture trends, thereby achieving more accurate sea ice risk prediction under complex environmental conditions, so as to help users take preventive measures in a timely manner and reduce potential risks in marine activities. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 A flow chart of a sea ice risk prediction method provided in Embodiment 1 of the present invention;

[0034] Figure 2 A schematic diagram of the structure of a sea ice risk prediction device provided in Embodiment 2 of the present invention;

[0035] Figure 3 This is a schematic diagram of the structure of a computer device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0036] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0037] It should be mentioned before discussing the exemplary embodiments in more detail that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the steps as sequential processes, many of the steps therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the steps can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the accompanying drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0038] Embodiment 1

[0039] Figure 1 A flow chart of a sea ice risk prediction method provided in the first embodiment of the present invention. This embodiment is applicable to situations where sea ice risk needs to be predicted in scenarios such as shipping safety assurance, marine oil and gas exploitation, offshore wind power, and offshore photovoltaics. In marine shipping, especially in the Arctic route and other cold waters, sea ice risk will directly threaten the safety of ship navigation. By real-time analysis and prediction of the possibility of sea ice formation, it can help ships choose the best route, avoid sea ice collision accidents, ensure shipping safety, and help ships plan shipping routes to reduce energy consumption and operating costs. When operating in cold waters, offshore oil and gas platforms are also faced with the risks brought by sea ice. The sudden appearance of sea ice may have a significant impact on the structural stability and operational safety of the platform. By predicting the sea ice risk and issuing an alarm in advance, the platform can be provided with advance response time so as to take protective measures, such as stopping exploitation, evacuating personnel, and strengthening structural reinforcement. Offshore wind farms and other marine engineering facilities (such as photovoltaic power stations) are also susceptible to sea ice. By predicting the upcoming sea ice in advance, timely maintenance suggestions are provided to operation and maintenance personnel, such as preventing wind turbines or photovoltaic panels from being damaged, thereby reducing facility downtime and maintenance costs. In addition, through long-term data analysis, a scientific basis can be provided for the site selection of marine engineering construction to reduce the impact of sea ice risk. The method can be executed by the sea ice risk prediction device provided by the embodiment of the present invention, which can be implemented by hardware and / or software and can generally be integrated into a computer device. Figure 1 As shown, the specific steps include:

[0040] S11. Obtain historical environmental data and historical sea ice monitoring data.

[0041] S12. Constructing a linear regression relationship model between the target environmental factor and the sea ice data, and a time series analysis model for the sea ice data based on the historical environmental data and the historical sea ice monitoring data.

[0042] S13, fusing the linear regression relationship model and the time series analysis model to obtain a hybrid model.

[0043] S14. Acquire real-time environmental data, and use the hybrid model to generate sea ice prediction data.

[0044] Specifically, historical environmental data may include meteorological data (such as surface air temperature, surface air pressure, surface wind speed and direction, surface relative humidity, irradiance, cloud cover, etc.) and ocean data (such as wave height, wave direction, tide, sea surface temperature, seawater salinity, seawater density, current velocity, current direction, wave height, etc.), and historical sea ice monitoring data may include sea ice thickness, sea ice coverage, etc. After obtaining the historical environmental data and historical sea ice monitoring data, the data can be standardized first to ensure that all data are in the same range and eliminate the impact of scale differences between different features to facilitate model processing. Specifically, the maximum-minimum normalization or Z-score normalization method can be used to standardize the obtained data. For the maximum-minimum normalization method, the normalization formula can be used for each feature x in the data set: Processing, where x is the original value, x norm is the normalized value, x min and x max are the minimum and maximum values ​​of feature x, respectively. The normalized data range is usually between [0,1], and can also be mapped to [-1,1]. For the Z-score normalization method, the normalization formula can be used: Processing is performed, where μ is the mean of feature x and σ is the standard deviation of feature x.

[0045] Based on the historical data obtained, on the one hand, the linear regression technology can be used to model the linear relationship between the target environmental factors (such as surface air temperature, surface air pressure, wind speed, etc.) and sea ice data. The formula is: y = beta0 + beta1x1 + beta2x2 + ... + beta n x n +epsilon, where y is the predicted variable (such as sea ice thickness), x i represents the target environmental factor, beta iRepresents the regression coefficient, and epsilon represents the error term. Linear regression is a statistical method used to explore the relationship between a dependent variable and one or more independent variables. The value of the target variable can be predicted through the relationship between the dependent variable (target variable) and one or more independent variables (feature variables). Linear regression can be used to explore the relationship between environmental factors related to sea ice risk, such as air temperature, sea water temperature, wind speed, ocean current speed and sea ice data, so as to clarify the causal relationship between these variables and sea ice changes in historical time periods. For example, when the temperature drops, the possibility of sea ice formation increases. This relationship can be quantified through linear regression, and the change in sea ice coverage for each degree of temperature drop can be obtained. In addition, in the sea ice risk warning system, multiple influencing factors need to be considered. Linear regression can help select features for these factors and find the factors that have the most significant impact on sea ice changes. These important features can be used to guide the construction and optimization of subsequent time series models, thereby improving the performance of the overall model. In addition, linear regression can not only predict the trend of sea ice formation, but also help evaluate the accuracy of the model through residual analysis. For example, by analyzing the residuals of the model, it can be determined whether the model has systematic deviations, which is very important for subsequent model adjustment and optimization. In practical applications, residual analysis can also help identify nonlinear features, which can also be supplemented by other methods. Modeling. By introducing ocean and meteorological factors closely related to sea ice changes into the modeling process, the ability to describe complex multivariate interactions is enhanced.

[0046] Optionally, before constructing a linear regression relationship model between the target environmental factor and the sea ice data based on the historical environmental data and the historical sea ice monitoring data, it also includes: using a Bayesian network or Granger causality analysis method to select the target environmental factor that has an influence on the sea ice data from various environmental factors in the environmental data.

[0047] Specifically, before constructing the linear regression relationship model, the key environmental factors closely related to sea ice changes can be analyzed based on historical data as target environmental factors, so as to construct a linear regression relationship model between the target environmental factors and sea ice data. This can be achieved by using Bayesian networks or Granger causality analysis methods. Bayesian networks explore the causal relationship between variables by constructing causal graphs, while Granger causality analysis methods are used to determine the extent to which a variable can predict another variable. Through causal analysis, the features that have the greatest impact on sea ice can be selected, so that the constructed linear regression relationship model can more accurately describe the relationship between environmental factors and sea ice data.

[0048] Based on the historical data obtained, on the other hand, time series analysis techniques can be used to model time series analysis models of sea ice data. Specifically, any time series analysis algorithm can be used. Time series analysis is a statistical method that is very suitable for time data. These data have a certain time order and may show periodicity, trend and random volatility. In the process of time series modeling, physical factors (such as the impact of temperature and tide on sea ice) can be combined for modeling to ensure that the output of the model conforms to known physical laws. Specifically, by adding penalty terms based on physical constraints to the time series analysis model, the physical rationality of the prediction results can be limited and the interpretability of the model can be improved.

[0049] Optionally, constructing a linear regression relationship model between the target environmental factor and the sea ice data, and a time series analysis model for the sea ice data based on the historical environmental data and the historical sea ice monitoring data includes: determining whether the historical sea ice monitoring data is a stationary time series; if so, modeling the time series analysis model based on an autoregressive moving average model; if not, performing differential processing on the historical sea ice monitoring data to convert it into a stationary time series, and modeling the time series analysis model based on an autoregressive difference moving average model.

[0050] The Autoregressive Moving Average (ARMA) model is used to describe the stable part of the time series data. It consists of an autoregressive (AR) part and a moving average (MA) part. The autoregressive part is used to establish the relationship between the current data point and the previous data points to capture the autocorrelation in the time series. The moving average part is used to model the random fluctuations in the data. The formula is: t =phi1y t-1 +phi2y t-2 +…+phi p y t-p +theta1epsilon t-1 +theta2epsilon t-2 +…+theta q epsilon t-q , where phi i represents the autoregressive coefficient, theta i Represents the moving average coefficient, epsilon tRepresents the error term. The ARMA model is suitable for stationary time series and can capture the stable changes in sea ice thickness and coverage over time. For example, the change in sea ice thickness usually shows a linear relationship with the thickness at previous time points, which can be modeled by the AR part, while the volatility of weather can be handled by the MA part.

[0051] For non-stationary time series with trends, they can be converted into stationary time series using differential processing, and then modeled using the Autoregressive Integrated Moving Average (ARIMA) model. The ARIMA model adds a differential operation to the ARMA model to handle long-term trends in time series. The formation and melting of sea ice are often affected by seasonal changes, so they show long-term trends. Through the ARIMA model, this trend effect can be removed to obtain a stable prediction model. For example, by differentiating the sea ice area over time series data, the impact of seasonal fluctuations can be effectively eliminated, making the prediction more accurate.

[0052] The stationarity of a time series refers to the fact that its statistical characteristics (mean, variance, autocorrelation structure, etc.) are constant over time. The stationarity of historical sea ice monitoring data can be determined by quantitative methods such as unit root test, KPSS test, variance stability test, combined with qualitative methods such as autocorrelation graph and time series graph. In the time series graph, observe the trend and fluctuation of the time series. If the mean or variance of the series changes significantly over time, it can be considered non-stationary. In the autocorrelation graph (ACF), check the autocorrelation function of the series. If the value of the autocorrelation function decreases sharply over time (generally decays to zero within a few steps), the series may be stationary. If the autocorrelation function decays slowly or changes periodically, the series may be non-stationary. The unit root test (Augmented Dickey-Fuller Test, ADF) is a commonly used statistical test to test whether a sequence has a unit root. If a sequence has a unit root, it means that it is non-stationary. The null hypothesis is that the sequence is non-stationary, that is, there is a unit root. If the P value is less than the significance level (usually 0.05), the null hypothesis is rejected and the sequence is considered to be stationary. The KPSS test (Kwiatkowski-Phillips-Schmidt-Shin Test) is used to test whether the mean of the sequence is stationary. In contrast to the ADF test, the KPSS test null hypothesis is that the sequence is stationary. If the P value is less than the significance level (usually 0.05), the null hypothesis is rejected and the sequence is considered to be non-stationary. The variance stability test observes whether the variance of the sequence is stable, such as calculating the local mean and variance through a sliding window. If the change is large, it may be non-stationary.

[0053] After determining that the historical sea ice monitoring data is a non-stationary time series, differential processing methods such as first difference and second difference can be used, depending on the degree of non-stationarity of the series. Specifically, when there is a linear trend in the time series, such as when the mean of the data rises or falls slowly over time, a single difference is usually sufficient to eliminate the trend and make the series stable. It can also be determined by the ADF test results. If the null hypothesis can be rejected after a single difference (the P value is less than the significance level), it means that the series has become stable, that is, a single difference can be used. The formula for a single difference is: ’ t =y t -y t-1 , where y t Represents the value of the time series at time t. When the time series shows a nonlinear trend, such as a parabolic trend, or the non-stationarity is not completely eliminated after the first difference, a second difference is usually required. It can also be judged by the ADF test results. If the sequence after the first difference still cannot pass the stationarity test (the P value is still high), a second difference can be performed to further eliminate the trend and make the series stable. The formula for the second difference is: ” t =y ’ t -y ’ t-1 =(y t -y t-1 )-(y t-1 -y t-2 )=y t -2y t-1 +y t-2 By effectively judging the stationarity of historical sea ice monitoring data and selecting an appropriate number of differences for preprocessing, it is easier to apply the ARIMA model for modeling in the future.

[0054] After completing the construction of the linear regression relationship model and the time series analysis model, the linear regression relationship model and the time series analysis model are fused to obtain a hybrid model. The linear regression relationship model is used to capture the long-term linear trend of meteorological conditions on sea ice formation, and can better explain the overall impact of meteorological conditions on sea ice formation, such as the impact of factors such as temperature and wind speed on sea ice formation and melting. The time series analysis model can capture the dynamic change characteristics of local fluctuations and randomness in sea ice data, especially when there are significant fluctuations and periodic changes, it can better describe the subtle changes in sea ice thickness and coverage. The fused hybrid model combines the explanatory power of linear regression and the dynamic capture ability of time series analysis, which can comprehensively describe the global trend and local fluctuations of sea ice changes, so that the model can achieve higher prediction accuracy in the face of complex sea ice change processes, especially when seasonal transitions or extreme weather occurs, which can significantly improve the prediction accuracy and stability. At the same time, in the sea ice risk warning, different environmental factors may have different characteristics. For example, seawater temperature may show a linear trend, while wind speed may show significant periodic fluctuations. Through the hybrid model, the linear regression part can be used for features with obvious linear relationships, while the time series analysis focuses on features with strong time dependence or periodicity, thereby achieving flexible modeling of different types of data. For example, the linear regression relationship model can reveal the impact of long-term changes in temperature and wind speed on the sea ice area, and the time series analysis model can make detailed predictions on fluctuations in local time periods on this basis, and then form a comprehensive judgment on future sea ice risks.

[0055] Optionally, the linear regression relationship model and the time series analysis model are fused to obtain a hybrid model, including: using a particle swarm optimization algorithm or a genetic algorithm to search for the optimal weight combination between the linear regression relationship model and the time series analysis model in the hybrid model. Specifically, an adaptive fusion mechanism can be used to adjust the weight combination of the linear regression and time series analysis parts according to the dynamic characteristics of the environmental data. Specifically, a particle swarm optimization algorithm (PSO) or a genetic algorithm (GA) can be used to search for the optimal weight combination, so that the hybrid model can have the best performance in the current data scenario, and as the environment changes, adaptive adjustments can also be made to ensure the best prediction performance in different scenarios, so that the model can flexibly adapt to environmental characteristics and capture important dynamic changes in different scenarios. At the same time, it also helps to improve the computational efficiency of the model, especially in a rapidly changing environment. It can still maintain efficient real-time warning capabilities to meet the real-time warning needs in scenarios that require rapid response, such as shipping and offshore oil and gas production. Furthermore, during the model fusion process, real-time causal analysis can be performed based on real-time environmental data, and the selection of target environmental factors and the combination of model weights can be dynamically adjusted to ensure that the model responds more sensitively to sea ice changes.

[0056] Optionally, a penalty term of a preset physical constraint is provided in the loss function of the hybrid model. Specifically, during the fusion process, the constraints of physical laws in oceanography and meteorology are introduced into the loss function to ensure that the output of the hybrid model conforms to the physical meaning. Exemplarily, the thickness of sea ice is restricted based on physical laws to ensure that the prediction results conform to the constraints of the actual environment, or the generation and melting rates of sea ice must be restricted within a reasonable range to ensure the physical consistency of the prediction. These physical constraints enable the model to provide reliable prediction results even under extreme weather conditions, ensuring that the prediction results have practical operational significance.

[0057] The system can collect environmental data in real time. After obtaining the hybrid model, it can be combined with the current real-time environmental data and use the hybrid model for prediction to obtain the current sea ice prediction data, such as sea ice thickness, coverage, etc. It can further determine the trend of sea ice changes and generate risk warning signals, etc.

[0058] Based on the above technical solution, optionally, after acquiring the real-time environmental data and using the hybrid model to generate sea ice prediction data, it also includes: acquiring real-time sea ice monitoring data corresponding to the sea ice prediction data; using an online learning mechanism to update the hybrid model online in combination with the real-time sea ice monitoring data to redefine the target environmental factors and the weight combination between the linear regression relationship model and the time series analysis model.

[0059] Specifically, the constructed hybrid model can be used for real-time prediction. Before each prediction, the latest real-time sea ice monitoring data can be obtained, and the parameters of the hybrid model can be updated online in combination with the latest real-time sea ice monitoring data input dynamically, so as to dynamically adapt to environmental changes and ensure that the model can quickly adjust in a changing environment, maintain sensitivity to the dynamic changes of sea ice, and provide accurate early warning signals in a timely manner. The online update process can also introduce physical constraints as penalty terms in the loss function to ensure that the prediction results of the model conform to physical laws. Among them, in view of the fact that sea ice changes are affected by multiple environmental variables, a dynamic causal analysis mechanism can be introduced. Specifically, Bayesian networks or Granger causal analysis methods can be used to construct a causal relationship network, evaluate the effects of different environmental factors on sea ice changes in real time, dynamically select the environmental factors that have the greatest impact on sea ice, and ensure that the linear regression relationship model can use the most influential features in real time in different environments, thereby enhancing the multivariate processing capability of the model and further enhancing the interpretability and adaptability of the model to complex environments. This interpretability is crucial for the decision-making process in scientific research, policy making, and emergency response, because it allows users to better understand the logic behind the prediction results, so that the prediction is no longer a "black box" operation, but has a clear causal explanation. For example, the model can dynamically indicate how factors such as temperature or wind speed affect sea ice coverage in a specific time period. For time series analysis models, such as ARIMA models, they can also be updated online based on the latest real-time sea ice monitoring data so that the latest data can be used to predict the sea ice data at the next moment. Subsequently, the latest hybrid model can be obtained based on the fusion of the updated linear regression relationship model and the time series analysis model. The fusion process can use the above-mentioned adaptive fusion mechanism to determine the optimal weight combination of the model in real time. By combining the above-mentioned online update mechanism, the model can cope with complex and changeable climate changes and ensure the accuracy of real-time warnings, especially in the face of sudden weather changes, and can still provide timely and reliable warning signals.

[0060] Through the organic combination of the above-mentioned technologies, a comprehensive, efficient and scientific prediction of sea ice changes has been achieved, providing a better solution for marine safety and giving the solution significant competitiveness in the field of sea ice risk warning. Especially in the face of complex and changeable meteorological conditions and high real-time requirements, the model can provide accurate predictions, rapid responses, reliable outputs and profound explanations, providing solid technical support for marine safety and shipping. It is not only suitable for operational decision-making, but also has application value in scientific research and policy making.

[0061] The technical solution provided by the embodiment of the present invention first obtains historical environmental data and historical sea ice monitoring data, then constructs a linear regression relationship model between the target environmental factor and the sea ice data, and a time series analysis model of the sea ice data based on the obtained historical environmental data and historical sea ice monitoring data, and then fuses the linear regression relationship model and the time series analysis model to obtain a hybrid model, and then obtains real-time environmental data, and uses the hybrid model to generate sea ice prediction data. By combining the ability of linear regression to explain influencing factors and the dynamic grasp of historical data by the time series analysis model, the laws of sea ice generation and change can be better described. By introducing a hybrid model, the advantages of different models can be effectively integrated to improve the prediction accuracy of sea ice risks and the ability to capture trends, thereby achieving more accurate sea ice risk prediction under complex environmental conditions, so as to help users take preventive measures in a timely manner and reduce potential risks in marine activities.

[0062] Embodiment 2

[0063] Figure 2 This is a schematic diagram of the structure of the sea ice risk prediction device provided in the second embodiment of the present invention. The device can be implemented by hardware and / or software, and can generally be integrated into a computer device to execute the sea ice risk prediction method provided in any embodiment of the present invention. Figure 2 As shown, the device comprises:

[0064] A data acquisition module 21 is used to acquire historical environmental data and historical sea ice monitoring data;

[0065] A model building module 22, for building a linear regression relationship model between target environmental factors and sea ice data, and a time series analysis model for sea ice data based on the historical environmental data and the historical sea ice monitoring data;

[0066] A model fusion module 23, used to fuse the linear regression relationship model and the time series analysis model to obtain a hybrid model;

[0067] The real-time prediction module 24 is used to obtain real-time environmental data and generate sea ice prediction data using the hybrid model.

[0068] The technical solution provided by the embodiment of the present invention first obtains historical environmental data and historical sea ice monitoring data, then constructs a linear regression relationship model between the target environmental factor and the sea ice data, and a time series analysis model of the sea ice data based on the obtained historical environmental data and historical sea ice monitoring data, and then fuses the linear regression relationship model and the time series analysis model to obtain a hybrid model, and then obtains real-time environmental data, and uses the hybrid model to generate sea ice prediction data. By combining the ability of linear regression to explain influencing factors and the dynamic grasp of historical data by the time series analysis model, the laws of sea ice generation and change can be better described. By introducing a hybrid model, the advantages of different models can be effectively integrated to improve the prediction accuracy of sea ice risks and the ability to capture trends, thereby achieving more accurate sea ice risk prediction under complex environmental conditions, so as to help users take preventive measures in a timely manner and reduce potential risks in marine activities.

[0069] On the basis of the above technical solution, optionally, the device further includes:

[0070] The target environmental factor selection module is used to select the target environmental factor that has an influence on the sea ice data from various environmental factors in the environmental data using a Bayesian network or Granger causality analysis method before constructing a linear regression relationship model between the target environmental factor and the sea ice data based on the historical environmental data and the historical sea ice monitoring data.

[0071] On the basis of the above technical solution, optionally, the model building module 22 is specifically used for:

[0072] Determining whether the historical sea ice monitoring data is a stationary time series;

[0073] If so, the time series analysis model is obtained based on the autoregressive moving average model;

[0074] If not, the historical sea ice monitoring data is differentially processed to be converted into a stationary time series, and the time series analysis model is obtained based on an autoregressive difference moving average model.

[0075] On the basis of the above technical solution, optionally, the model fusion module 23 is specifically used for:

[0076] A particle swarm optimization algorithm or a genetic algorithm is used to search for the optimal weight combination between the linear regression relationship model and the time series analysis model in the hybrid model.

[0077] Based on the above technical solution, optionally, a penalty term of a preset physical constraint is set in the loss function of the hybrid model.

[0078] On the basis of the above technical solution, optionally, the device further includes:

[0079] A real-time data acquisition module, configured to acquire real-time sea ice monitoring data corresponding to the sea ice prediction data after acquiring the real-time environmental data and generating the sea ice prediction data using the hybrid model;

[0080] The model online updating module is used to adopt an online learning mechanism to update the hybrid model online in combination with the real-time sea ice monitoring data to redefine the target environmental factors and the weight combination between the linear regression relationship model and the time series analysis model.

[0081] The sea ice risk prediction device provided in the embodiment of the present invention can execute the sea ice risk prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0082] It is worth noting that in the above-mentioned embodiment of the sea ice risk prediction device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0083] Embodiment 3

[0084] Figure 3 The schematic diagram of the structure of the computer device provided for the third embodiment of the present invention shows a block diagram of an exemplary computer device suitable for implementing the implementation mode of the present invention. Figure 3 The computer device shown is only an example and should not limit the functions and scope of use of the embodiments of the present invention. Figure 3 As shown, the computer device includes a processor 31, a memory 32, an input device 33 and an output device 34; the number of processors 31 in the computer device can be one or more. Figure 3 Taking a processor 31 as an example, the processor 31, memory 32, input device 33 and output device 34 in the computer device can be connected through a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0085] The memory 32, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the sea ice risk prediction method in the embodiment of the present invention (for example, the data acquisition module 21, the model building module 22, the model fusion module 23 and the real-time prediction module 24 in the sea ice risk prediction device). The processor 31 executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory 32, that is, realizes the above-mentioned sea ice risk prediction method.

[0086] The memory 32 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 32 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 32 may further include a memory remotely arranged relative to the processor 31, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0087] The input device 33 can be used to obtain historical environmental data and historical sea ice monitoring data in real time, and to generate key signal inputs related to user settings and function control of the computer device, etc. The output device 34 can be used to provide sea ice prediction data, etc. to the user.

[0088] Embodiment 4

[0089] Embodiment 4 of the present invention further provides a storage medium comprising computer executable instructions, wherein the computer executable instructions are used to execute a sea ice risk prediction method when executed by a computer processor, the method comprising:

[0090] Obtain historical environmental data and historical sea ice monitoring data;

[0091] Constructing a linear regression relationship model between target environmental factors and sea ice data, and a time series analysis model for sea ice data based on the historical environmental data and the historical sea ice monitoring data;

[0092] The linear regression relationship model and the time series analysis model are integrated to obtain a hybrid model;

[0093] Real-time environmental data is acquired and sea ice prediction data is generated using the hybrid model.

[0094] The storage medium may be any of various types of memory devices or storage devices. The term "storage medium" is intended to include: installation media, such as CD-ROM, floppy disk or tape device; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (such as hard disk or optical storage); registers or other similar types of memory elements, etc. The storage medium may also include other types of memory or combinations thereof. In addition, the storage medium may be located in the computer system in which the program is executed, or may be located in a different second computer system, which is connected to the computer system via a network (such as the Internet). The second computer system may provide program instructions to the computer for execution. The term "storage medium" may include two or more storage media that may reside in different locations (e.g., in different computer systems connected via a network). The storage medium may store program instructions (e.g., embodied as a computer program) that may be executed by one or more processors.

[0095] Of course, the storage medium containing computer executable instructions provided in an embodiment of the present invention, whose computer executable instructions are not limited to the method operations described above, can also execute related operations in the sea ice risk prediction method provided in any embodiment of the present invention.

[0096] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0097] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0098] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0099] Embodiment 5

[0100] Embodiment 5 of the present invention also provides a computer program product, which includes a computer program (also referred to as code, instruction), which can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it is used to execute the sea ice risk prediction method provided in any of the above embodiments, and has the corresponding beneficial effects of the execution method.

[0101] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A sea ice risk prediction method, characterized in that: include: Obtain historical environmental data and historical sea ice monitoring data; Constructing a linear regression relationship model between target environmental factors and sea ice data, and a time series analysis model for sea ice data based on the historical environmental data and the historical sea ice monitoring data; The linear regression relationship model and the time series analysis model are integrated to obtain a hybrid model; Real-time environmental data is acquired and sea ice prediction data is generated using the hybrid model.

2. The sea ice risk prediction method according to claim 1, characterized in that: Before constructing a linear regression relationship model between the target environmental factor and the sea ice data according to the historical environmental data and the historical sea ice monitoring data, the method further includes: The target environmental factor having an influence on the sea ice data is selected from various environmental factors in the environmental data by using a Bayesian network or Granger causality analysis method.

3. The sea ice risk prediction method according to claim 1, characterized in that: The linear regression relationship model between the target environmental factor and the sea ice data and the time series analysis model of the sea ice data are constructed based on the historical environmental data and the historical sea ice monitoring data, including: Determining whether the historical sea ice monitoring data is a stationary time series; If so, the time series analysis model is obtained based on the autoregressive moving average model; If not, the historical sea ice monitoring data is differentially processed to be converted into a stationary time series, and the time series analysis model is obtained based on an autoregressive difference moving average model.

4. The sea ice risk prediction method according to claim 1, characterized in that: The step of fusing the linear regression relationship model and the time series analysis model to obtain a hybrid model includes: A particle swarm optimization algorithm or a genetic algorithm is used to search for the optimal weight combination between the linear regression relationship model and the time series analysis model in the hybrid model.

5. The sea ice risk prediction method according to claim 1, characterized in that: A penalty term of a preset physical constraint is set in the loss function of the hybrid model.

6. The sea ice risk prediction method according to claim 1, characterized in that: After acquiring the real-time environmental data and generating sea ice prediction data using the hybrid model, the method further includes: Acquire real-time sea ice monitoring data corresponding to the sea ice prediction data; An online learning mechanism is adopted to update the hybrid model online in combination with the real-time sea ice monitoring data to redefine the target environmental factors and the weight combination between the linear regression relationship model and the time series analysis model.

7. A sea ice risk prediction device, characterized in that: include: A data acquisition module, used to obtain historical environmental data and historical sea ice monitoring data; A model building module, used to build a linear regression relationship model between the target environmental factor and the sea ice data, and a time series analysis model for the sea ice data based on the historical environmental data and the historical sea ice monitoring data; A model fusion module, used to fuse the linear regression relationship model and the time series analysis model to obtain a hybrid model; The real-time prediction module is used to obtain real-time environmental data and generate sea ice prediction data using the hybrid model.

8. A computer device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the sea ice risk prediction method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the sea ice risk prediction method as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the computer program implements the sea ice risk prediction method as described in any one of claims 1 to 6.

Images