Ship software operation and maintenance method and device, electronic equipment and storage medium

Abstract

The invention discloses a ship software operation and maintenance method and device, electronic equipment and a storage medium. According to the specific scheme, first data of a target ship in multiple dimensions within a preset duration before a ship software fault moment is obtained, wherein the first data in the multiple dimensions comprises at least two of meteorological data, equipment state data, equipment environment data, geographic position data and operation behavior data; processing the first data based on a random forest regression model used for determining ship faults to obtain a fault prediction result, the fault prediction result comprising a fault evaluation attribute corresponding to the first data under each dimension, and the random forest regression model being a model deployed in the target ship; based on the fault prediction result and the first data, ship software operation and maintenance information is generated, and software maintenance processing is carried out on ship software of the target ship based on the ship software operation and maintenance information. According to the method, the fault reason of the ship software is quickly and accurately positioned, and the operation and maintenance efficiency of the ship software is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method, device, electronic device and storage medium for ship software operation and maintenance. Background Art

[0002] With the development of intelligent ship technology, more and more software is running on ships. These software include data acquisition software, ship optimization software, video surveillance software, etc. Due to the complex and changeable environment of ships, these software are extremely vulnerable to interference from factors such as bad weather, complex routes or equipment aging, resulting in software system crashes, hangs or unstable operation.

[0003] Based on the above situation, corresponding operation and maintenance personnel usually remotely maintain ship software. However, manual operation depends on the technical experience of operation and maintenance personnel, and it may be difficult to accurately and quickly determine the cause of specific faults, so it is difficult to perform targeted operation and maintenance on ship software. In addition, the satellite signal used to support the ship network is easily affected by sea area changes, and may not provide a stable network signal, resulting in operation and maintenance personnel often being unable to remotely maintain ship software. Based on this, problems such as low efficiency of ship software operation and maintenance and difficulty in locating the cause of faults will occur. Summary of the Invention

[0004] The present invention provides a method, device, electronic device and storage medium for ship software operation and maintenance, which realizes the rapid and accurate positioning of the cause of ship software faults and improves the operation and maintenance efficiency of ship software.

[0005] According to an aspect of the present invention, a method for ship software operation and maintenance is provided. The method includes:

[0006] Obtain first data in multiple dimensions within a preset time period before the ship software failure moment of the target ship, where the first data in multiple dimensions includes at least two of: meteorological data corresponding to the target ship, equipment status data of the target ship, environment data where the equipment of the target ship is located, geographical location data of the target ship, and operation behavior data of the target ship;

[0007] Process the first data based on a pre-trained random forest regression model for determining ship faults to obtain a fault prediction result, where the fault prediction result includes a fault evaluation attribute corresponding to the first data in each dimension, and the random forest regression model is a model deployed in the target ship;

[0008] Generate ship software operation and maintenance information based on the fault prediction result and the first data, so as to perform software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information.

[0009] According to another aspect of the present invention, there is provided a ship software operation and maintenance device, which includes:

[0010] A data acquisition module, configured to acquire first data in multiple dimensions within a preset time period before the moment of ship software failure of a target ship. Among them, the first data in multiple dimensions includes at least two of the following: meteorological data corresponding to the target ship, equipment status data of the target ship, environmental data where the equipment of the target ship is located, geographical location data of the target ship, and operation behavior data of the target ship;

[0011] A fault determination module, configured to process the first data based on a pre-trained random forest regression model for determining ship faults to obtain a fault prediction result. Among them, the fault prediction result includes a fault evaluation attribute corresponding to the first data in each dimension, and the random forest regression model is a model deployed in the target ship;

[0012] A ship software maintenance module, configured to generate ship software operation and maintenance information based on the fault prediction result and the first data, so as to perform software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information.

[0013] According to another aspect of the present invention, there is provided an electronic device, which includes:

[0014] At least one processor; and

[0015] A memory communicatively connected to at least one processor; wherein,

[0016] The memory stores a computer program executable by at least one processor. The computer program is executed by at least one processor so that at least one processor can execute the ship software operation and maintenance method of any embodiment of the present invention.

[0017] According to another aspect of the present invention, there is provided a computer-readable storage medium storing computer instructions for causing a processor to implement the ship software operation and maintenance method of any embodiment of the present invention when executed.

[0018] According to another aspect of the present invention, there is provided a computer program product including a computer program, characterized in that the computer program implements the ship software operation and maintenance method of any embodiment of the present invention when executed by a processor.

[0019] The technical solution of the embodiment of the present invention obtains the first data in multiple dimensions within a preset time period before the moment of ship software failure for the target ship, and processes the first data through a pre-trained random forest regression model deployed on the target ship for determining ship faults to determine the fault evaluation attributes corresponding to the first data in each dimension, that is, obtain the fault prediction result, and determine the probability that the first data in each dimension causes the ship software of the target ship to fail. According to the fault prediction result, the factors affecting the normal operation of the ship software can be determined, solving the problems in the prior art that it is difficult to determine the specific fault cause and the manual operation and maintenance cost is too high caused by manual remote operation and maintenance of the ship software. Through the fault prediction result and the first data, ship software operation and maintenance information is generated to perform software maintenance on the ship software of the target ship based on the ship software operation and maintenance information. The present invention solves the problems such as high operation and maintenance cost, low ship software operation and maintenance efficiency, and difficult fault cause location caused by manual remote operation and maintenance of the ship software. By analyzing the first data in multiple dimensions to determine the fault prediction result, the accuracy of fault cause location is guaranteed, and more comprehensive software maintenance processing can be provided for the ship software of the target ship. Through the random forest regression model deployed on the target ship for fault detection, it is realized that the ship software anomaly can be automatically processed without manual remote operation and maintenance of the ship software, improving the operation and maintenance efficiency of the ship software, and ensuring the stability and safety of the operation of the ship software of the target ship in the case where there is no satellite network communication and the operation and maintenance personnel cannot remotely operate and maintain the ship software.

[0020] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Brief Description of the Drawings

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0022] Figure 1 It is a flowchart of a ship software operation and maintenance method provided by an embodiment of the present invention;

[0023] Figure 2 It is a structural example diagram of a ship software operation and maintenance system provided by an embodiment of the present invention;

[0024] Figure 3 It is a flowchart of a ship software operation and maintenance method provided by an embodiment of the present invention;

[0025] Figure 4 It is a schematic structural diagram of a ship software operation and maintenance device provided by an embodiment of the present invention;

[0026] Figure 5 It is a schematic structural diagram of an electronic device for implementing the ship software operation and maintenance method of the embodiment of the present invention. Detailed implementation manners

[0027] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0029] Embodiment 1

[0030] Figure 1 It is a flowchart of a ship software operation and maintenance method provided by Embodiment 1 of the present invention. This embodiment is applicable to processing first data in multiple dimensions through a random forest regression model deployed on a target ship for determining ship faults, so as to quickly, accurately and comprehensively determine the reasons for ship software faults, and then automatically perform operation and maintenance processing on the ship software. This method can be executed by a ship software operation and maintenance device, which can be implemented in the form of hardware and / or software, and the ship software operation and maintenance device can be configured in an electronic device such as a mobile phone, a computer or a server. As Figure 1 shown, the method includes:

[0031] S110. Obtain first data in multiple dimensions within a preset time period before the ship software fails for the target ship.

[0032] Among them, the first data in multiple dimensions includes at least two of the meteorological data corresponding to the target ship, the equipment status data of the target ship, the environment data where the equipment of the target ship is located, the geographical location data of the target ship, and the operation behavior data of the target ship. The target ship can be a ship that is currently sailing. A pre-trained random forest regression model for determining ship faults is deployed in the target ship to timely determine the cause of the ship software fault through the random forest regression model when a fault in the ship software is detected. The ship software can be understood as the software deployed in the target ship. For example, the ship software can be the system software corresponding to the target ship. The fault moment can be the moment when the ship software of the target ship has a fault. Since the fault of the ship software may be caused by one or more factors, first data in multiple dimensions can be obtained. The preset duration can be a period of time set in advance before the fault moment of the ship software.

[0033] The first data in multiple dimensions can include at least two of the meteorological data corresponding to the target ship, the equipment status data of the target ship, the environment data where the equipment of the target ship is located, the geographical location data of the target ship, and the operation behavior data of the target ship. Among them, the meteorological data can include data such as the wave height, wave direction, and wind level in the sea area where the target ship is located. The equipment status data can include data such as the usage rate of the central processing unit (CPU), memory usage rate, and disk usage rate corresponding to the ship software of the target ship. The environment data where the equipment is located can be understood as data such as the temperature and humidity of the environment where the target ship is located. The geographical location data of the target ship can be understood as the longitude and latitude information where the target ship is located. The operation behavior data can be understood as the operation behavior data generated when the staff in the target ship interacts with the ship software. Optionally, since there is corresponding data in multiple dimensions at each moment within the preset duration, in order to improve the calculation efficiency, the first data obtained can be mean data. For example, taking the preset duration as 10 minutes and the first data as the equipment status data, the average CPU usage rate within 10 minutes before the fault moment of the ship software can be used as the first data in the current dimension.

[0034] Specifically, when it is detected that the ship software of the target ship has situations such as system suspension or crash, it is determined that the ship software of the target ship has a fault. Obtain the first data in multiple dimensions of the target ship within the preset duration before the fault moment of the ship software. Among them, the first data in multiple dimensions includes at least two of the meteorological data corresponding to the target ship, the equipment status data of the target ship, the environment data where the equipment of the target ship is located, the geographical location data of the target ship, and the operation behavior data of the target ship, so as to provide data support for subsequent determination of the cause of the ship software fault and ship software operation and maintenance through the first data in multiple dimensions.

[0035] Exemplarily, referring to Figure 2 , the ship software operation and maintenance method can be implemented by a ship software operation and maintenance system. The ship software operation and maintenance system includes: a data receiving module, a data storage module, a data processing module, and a model training module deployed on the shore server, and a data acquisition module, an intelligent decision-making module, an automatic control module, and a feedback module deployed on the ship server of the target ship. Each module in the shore server is mainly used to construct and train a random forest regression model for determining ship faults based on the received data. The data acquisition module of the ship server in the target ship is used to acquire meteorological data, equipment status data, external environment data, and operation behavior data. Among them, the external environment data corresponds to the environment data of the equipment of the target ship mentioned above. The data acquired by the data acquisition module is transmitted to the shore server through satellite communication, so that the shore server constructs and trains a random forest regression model for determining ship faults based on the received data. It should be noted that after the random forest regression model for determining ship faults has been trained, the random forest regression model for determining ship faults can be packaged into an offline model prediction file and transmitted to the ship server and deployed in the intelligent decision-making module of the ship server. When a fault in the ship software of the target ship is detected, the first data in multiple dimensions within a preset time period before the ship software fault moment of the target ship can be obtained through the data acquisition module, and the first data in multiple dimensions is transmitted to the intelligent decision-making module to process the first data based on the random forest regression model for determining ship faults deployed in the intelligent decision-making module, and then determine the factors causing the ship software to fail.

[0036] In the embodiment of the present invention, the acquisition method of the first data may be: when a fault of the target ship is detected, acquiring the original data in multiple dimensions within a preset time period before the ship software fault moment of the target ship; performing data preprocessing on the original data to obtain the first data associated with the ship software fault of the target ship, where the data preprocessing includes at least one of data cleaning processing, outlier processing, and data normalization processing based on preset data screening conditions.

[0037] Among them, the original data can be understood as the initial data in multiple dimensions within a preset time period before the fault moment. There may be abnormal data, missing data, etc. in the original data. The original data in multiple dimensions includes at least two of the meteorological data corresponding to the target ship, the equipment status data of the target ship, the environmental data where the equipment of the target ship is located, the geographical location data of the target ship, and the operation behavior data of the target ship. Since there may be abnormal data, missing data, etc. in the original data, data preprocessing can be performed on the original data. The data preprocessing can include at least one of data cleaning processing, outlier processing, and data normalization processing based on preset data screening conditions. The preset data screening conditions can be data screening conditions set according to the preset time period. For example, if the preset time period is 10 minutes, the preset data screening conditions can be that if the time corresponding to the original data is within 10 minutes before the fault moment, the data is retained, and other data is removed. The outlier processing can be to remove the data in the original data that exceeds the corresponding preset data threshold. The data normalization processing can be to scale the original data according to a certain ratio so that it falls within a specific interval. The purpose of normalization is to eliminate the differences in dimension and value range between the original data, so that the data in different dimensions are comparable. Optionally, the Z-Score normalization algorithm can be used to make the data in different dimensions all within the value range of [0,1] to eliminate the scale differences between the data in different dimensions.

[0038] Specifically, when it is detected that there is a fault in the ship software of the target ship, at least two-dimensional original data such as the meteorological data, equipment status data, environmental data where the equipment is located, geographical location data, and operation behavior data of the ship software of the target ship within a preset time period before the fault moment can be obtained. The original data is screened according to the preset data screening conditions corresponding to the preset time period, and outlier processing and normalization processing can also be performed on the screened original data to obtain the first data corresponding to the original data. It should be noted that in order to improve the calculation efficiency, after normalizing each dimension of the original data to obtain the normalized data corresponding to each dimension, the normalized data of each dimension can be averaged to obtain the average data corresponding to each dimension, and this average data can be used as the first data.

[0039] S120. Process the first data based on the pre-trained random forest regression model for determining ship faults to obtain a fault prediction result, where the fault prediction result includes the fault evaluation attributes corresponding to the first data in each dimension, and the random forest regression model is a model deployed in the target ship.

[0040] Among them, the random forest regression model can be used to process the input first data, and through the output results of each decision tree, obtain the fault prediction result of the target ship. Optionally, decision tree algorithms such as gradient boosting decision tree, XGBoost or LightGBM can also be used to implement the fault prediction processing of the target ship. The fault prediction result includes the fault evaluation attributes corresponding to the first data in each dimension. The fault evaluation attribute can be used to represent the probability value of the first data in the current dimension causing a fault in the ship software. According to the fault prediction result, the probability value of the first data in each dimension causing a fault in the ship software can be determined.

[0041] Specifically, each decision tree of the pre-trained random forest regression model for determining ship faults deployed in the target ship is used to process the first data in multiple dimensions, and the fault evaluation attribute of the first data in each dimension causing a fault in the ship software is determined. And the fault evaluation attributes corresponding to multiple dimensions are determined as the fault prediction result, so as to perform maintenance processing on the ship software of the target ship according to the fault prediction result.

[0042] Exemplarily, in combination with the above example, see Figure 2 , according to the processing of the first data by the random forest regression model for determining ship faults deployed in the intelligent decision-making module, the fault evaluation attributes corresponding to the first data in each dimension are obtained, that is, the fault prediction result is obtained.

[0043] S130. Generate ship software operation and maintenance information based on the fault prediction result and the first data, so as to perform software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information.

[0044] Among them, the ship software operation and maintenance information can be prompt information for performing maintenance processing such as restarting the ship software or optimizing resource allocation.

[0045] Specifically, according to the fault evaluation attributes corresponding to the first data in each dimension in the fault prediction result and the corresponding preset attribute threshold, the dimension causing a fault in the ship software is determined, and the first data in this dimension is determined. According to the first data in this dimension and the fault evaluation attribute corresponding to this dimension in the fault prediction result, ship software operation and maintenance information is generated, so as to perform software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information.

[0046] Exemplarily, taking the first data in multiple dimensions as meteorological data and operation behavior data as an example, if the fault evaluation attribute corresponding to the meteorological data in the fault prediction result is greater than the corresponding preset attribute threshold, while the fault evaluation attribute corresponding to the operation behavior data is less than the corresponding preset attribute threshold, it can be determined that the ship software fault is caused by the meteorological data. According to the meteorological data and the fault evaluation attribute corresponding to the meteorological data, if the determined ship software operation and maintenance information is the information for restarting the ship software, then the ship software can be automatically restarted according to the ship software operation and maintenance information.

[0047] Exemplarily, referring to Figure 2 , combining the above example, after the intelligent decision-making module determines the fault prediction result, the intelligent decision-making module can generate ship software operation and maintenance information according to the fault prediction result and the first data. And send the ship software operation and maintenance information to the automatic control module, so that the automatic control module automatically executes corresponding software maintenance operations according to the ship software operation and maintenance information. For example, the software maintenance operation can be to restart the software service, optimize the resource configuration of the ship software, etc., to achieve the repair of the ship software fault and ensure the stable operation of the ship software.

[0048] Optionally, after performing software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information, the method further includes: updating the model parameters of the random forest regression model used to determine ship faults based on the operation and maintenance log corresponding to the ship software operation and maintenance information, to obtain an updated random forest regression model, so as to process the next first data based on the updated random forest regression model to obtain the corresponding fault prediction result.

[0049] Among them, the operation and maintenance log can be used to characterize the specific situation of performing maintenance processing on the ship software based on the ship software operation and maintenance information. Optionally, the operation and maintenance log is sent to the terminal device of the corresponding operation and maintenance personnel by means of text message, phone call or email, or the operation and maintenance log is displayed on the corresponding operation and maintenance monitoring interface, so as to facilitate the operation and maintenance personnel to trace the software maintenance processing process.

[0050] Specifically, during the process of automatically performing software maintenance on the ship software of the target ship based on the ship software operation and maintenance information, the software maintenance situation can be detected in real time, and after the ship software maintenance is completed, an operation and maintenance log corresponding to the ship software operation and maintenance information is generated. Optimize and update the model parameters of the pre-trained random forest regression model used to determine ship faults according to the operation and maintenance log, to obtain an updated random forest regression model, so as to process the next first data based on the updated random forest regression model to obtain the corresponding fault prediction result.

[0051] Optionally, during the process of real-time detection of software maintenance, if an abnormality is detected in the automatic maintenance of the ship software, a warning prompt message can be generated based on the software maintenance situation, and the warning prompt message can be sent to the terminal device of the corresponding operation and maintenance personnel by means of text message, phone call or email, or the warning prompt message can be displayed on the corresponding operation and maintenance monitoring interface, so as to facilitate the operation and maintenance personnel to maintain the ship software in a timely manner.

[0052] Exemplarily, in combination with the above example, refer to Figure 2 , the feedback module is used to detect the software maintenance situation of the ship software by the automatic control module in real time and generate an operation and maintenance log according to the software maintenance situation. The operation and maintenance log is sent to the terminal device of the corresponding operation and maintenance personnel of the target ship or the shore end by means of text message, phone call or email, so as to optimize and update the model parameters of the random forest regression model deployed in the intelligent decision-making module for determining ship faults according to the evaluation results of the operation and maintenance personnel on the operation and maintenance log, so as to ensure the accuracy of ship software operation and maintenance processing through a closed-loop feedback mechanism.

[0053] The technical solution of this embodiment obtains the first data in multiple dimensions within a preset time period before the ship software failure moment of the target ship, and processes the first data through a pre-trained random forest regression model for determining ship faults deployed on the target ship to determine the fault evaluation attributes corresponding to the first data in each dimension, that is, obtain the fault prediction result, and determine the probability that the first data in each dimension causes the ship software of the target ship to fail. According to the fault prediction result, the factors affecting the normal operation of the ship software can be determined, solving the problems in the prior art that it is difficult to determine the specific fault cause and the manual operation and maintenance cost is too high caused by manual remote operation and maintenance of the ship software. Through the fault prediction result and the first data, ship software operation and maintenance information is generated to perform software maintenance on the ship software of the target ship based on the ship software operation and maintenance information. The present invention solves the problems such as high operation and maintenance cost, low ship software operation and maintenance efficiency and difficult fault cause location caused by manual remote operation and maintenance of the ship software. By analyzing the first data in multiple dimensions and determining the fault prediction result, the accuracy of fault cause location is ensured, and more comprehensive software maintenance processing can be provided for the ship software of the target ship. Through the random forest regression model deployed on the target ship for fault detection, it is realized that the ship software anomaly can be automatically processed without manual remote operation and maintenance of the ship software, improving the operation and maintenance efficiency of the ship software and ensuring the stability and safety of the ship software operation of the target ship in the case where there is no satellite network communication and the operation and maintenance personnel cannot remotely operate and maintain the ship software.

[0054] Embodiment 2

[0055] Figure 3It is a flowchart of a ship software operation and maintenance method provided in the second embodiment of the present invention. On the basis of the above embodiment, before processing the first data using the pre-trained random forest regression model for determining ship faults, a random forest regression model for determining ship faults can be constructed and trained first. The specific implementation manner can refer to the technical solution of this embodiment. Among them, the same or corresponding technical terms as those in the above embodiment will not be described in detail here. As Figure 3 shown, the method includes:

[0056] S210. Obtain a plurality of original training samples, where the original training samples include original sample data and corresponding theoretical fault results, and the original sample data includes historical ship software fault data in multiple dimensions.

[0057] Among them, the original sample data may include: historical ship software fault data in multiple dimensions within a preset time period before the historical fault moment of the ship software of the target ship. For example, the historical ship software fault data in multiple dimensions may include at least two of the meteorological data corresponding to the target ship, the equipment status data of the target ship, the environmental data where the equipment of the target ship is located, the geographical location data of the target ship, and the operation behavior data of the target ship. Optionally, in order to ensure the accuracy of subsequent model training, the historical ship software fault data can be preprocessed to ensure that the original sample data does not contain abnormal or missing data. The theoretical fault result may include the theoretical fault probability value corresponding to the historical ship software fault data in each dimension. For example, if the fault of the historical ship software is caused by meteorological data, the fault probability value corresponding to the meteorological data in the theoretical fault result is 100%, and the fault probability values corresponding to the historical ship software fault data in other dimensions are 0%.

[0058] Specifically, before constructing and training the random forest regression model, a plurality of original training samples can be obtained first to construct and train a random forest regression model for determining ship faults based on the original training samples. In order to improve the accuracy of the model, as many and rich original training samples as possible can be obtained, historical ship software fault data in multiple dimensions within a preset time period at multiple historical fault moments are obtained, and the corresponding theoretical fault results at each historical fault moment are determined. The historical ship software fault data in multiple dimensions within a preset time period at each historical fault moment is used as the original sample data, so as to construct rich original training samples based on the above method.

[0059] Exemplarily, in combination with the above example, refer to Figure 2, the data receiving module of the shore server can obtain historical ship software fault data in multiple dimensions within a preset duration at multiple historical fault moments, as well as the corresponding theoretical fault results for each historical fault moment, and store these data in the data storage module using a distributed storage architecture to achieve efficient storage of data and subsequent rapid acquisition, ensuring the high availability and security of the data. The data processing module performs at least one data preprocessing such as data cleaning, outlier handling, and data normalization on the collected historical ship software fault data in multiple dimensions based on preset data screening conditions to obtain the original sample data. The original sample data and the corresponding theoretical fault results are used as the original training samples and transmitted to the model training module to construct and train a random forest regression model for determining ship faults based on the original training samples.

[0060] S220. Sample the original sample data of multiple original training samples according to a preset sampling algorithm to obtain the first sample data, and obtain the first training sample based on the first sample data and the corresponding theoretical fault result.

[0061] Among them, the preset sampling algorithm can be a sampling algorithm preset according to actual needs. Optionally, the preset sampling algorithm can be the Bootstrap sampling algorithm. The Bootstrap sampling algorithm is a resampling technique used to repeatedly sample with replacement from the original sample data of multiple original training samples to obtain the first sample data. The first training sample includes the first sample data and the theoretical fault result corresponding to the first sample data.

[0062] Specifically, sample the original sample data of multiple original training samples according to the preset sampling algorithm to obtain the first sample data. Determine the first training sample based on the first sample data and the corresponding theoretical fault result in the original training sample to construct a random forest regression model based on the first training sample.

[0063] S230. Based on the first training sample, use a preset decision tree algorithm to determine at least one decision tree to obtain the to-be-trained random forest regression model based on the at least one decision tree.

[0064] Among them, the preset decision tree algorithm can be a pre-set algorithm for constructing a decision tree.

[0065] Specifically, use the preset decision tree algorithm to process the first training sample to generate at least one decision tree, and construct the to-be-trained random forest regression model according to the at least one decision tree.

[0066] Exemplarily, the Bootstrap sampling algorithm is used as an example of the preset sampling algorithm for illustration. During the process of training and constructing the random forest regression model, the Bootstrap sampling algorithm is used to randomly extract the first sample data from the original training samples, and the theoretical fault results corresponding to the first sample data are determined. According to the first sample data and the theoretical fault results corresponding to the first sample data, the first training sample is determined, and based on the first training sample and the preset decision tree algorithm, at least one decision tree is generated to construct the random forest regression model to be trained based on at least one decision tree. It should be noted that in order to improve the efficiency of constructing the random forest regression model to be trained, at least one decision tree can be generated in parallel to improve the model construction efficiency through parallel computing.

[0067] S240. For at least one decision tree of the random forest regression model to be trained, the original sample data that has not been used to construct the decision tree is used as the second sample data, and based on the second sample data and the theoretical fault results corresponding to the second sample data, the second training sample is obtained.

[0068] Among them, the second sample data includes historical ship software fault data in multiple dimensions. The theoretical fault results of the second sample data are determined according to the theoretical fault results in the original training samples. The second training sample includes the second sample data and the theoretical fault results corresponding to the second sample data.

[0069] Specifically, for at least one decision tree of the random forest regression model to be trained, when constructing the decision tree using the sampled first sample data, there will be a part of the original sample data that is not used for each decision tree. Then, the original sample data that has not been used to construct the decision tree can be used as the second sample data, and based on the second sample data and the theoretical fault results in the original training samples corresponding to the second sample data, the second training sample is determined to train the random forest regression model to be trained based on the second training sample.

[0070] S250. The random forest regression model to be trained is trained based on the second training sample to obtain a trained random forest regression model, and the trained random forest regression model is determined as the random forest regression model for determining ship faults.

[0071] Specifically, the second sample data of the second training sample is input into the random forest regression model to be trained for training, so as to correct the model parameters of the random forest regression model to be trained through the output results of the random forest regression model to be trained and the theoretical fault results corresponding to the second sample data, and obtain the random forest regression model for determining ship faults.

[0072] In the implementation of the present invention, the specific manner of training the random forest regression model to be trained with the second training sample may be as follows: input the second sample data of the second training sample into at least one decision tree of the random forest regression model to be trained to obtain the first output result of each decision tree, where the first output result includes the fault evaluation attributes corresponding to the historical ship software fault data in each dimension; determine the first prediction result based on the first output result of each decision tree and the number of decision trees, and determine the first error data based on the first prediction result and the theoretical fault result corresponding to the second sample data; perform data augmentation processing on the second training sample to obtain a third training sample, where the third training sample includes the third sample data corresponding to the second sample data and the theoretical fault result corresponding to the third sample data; input the third sample data into at least one decision tree of the random forest regression model to be trained to obtain the second prediction result, and obtain the second error data according to the second prediction result and the theoretical fault result corresponding to the third sample data; correct the model parameters of the random forest regression model to be trained based on the first error data and the second error data to obtain the trained random forest regression model.

[0073] Among them, the second sample data includes historical ship software fault data in multiple dimensions. Correspondingly, the first output result includes the fault evaluation attributes corresponding to the historical ship software fault data in each dimension. The fault evaluation attribute is used to represent the predicted probability value of the historical ship software fault data causing the target ship to fail. The first prediction result can be used to represent the predicted probability value of the ship software of the target ship failing caused by the second sample data in each dimension output by the random forest regression model. The first prediction result can be the result obtained by performing a mean process on the first output result of each decision tree according to the number of decision trees. The first error data can be used to represent the degree of difference between the first prediction result and the theoretical fault result corresponding to the second sample data.

[0074] The data augmentation processing may be to randomly shuffle the second sample data in the second training sample or add noise to the second sample data in the second training sample. The third training sample is the second training sample after data augmentation processing. The third sample data in the third training sample is the second sample data after data augmentation processing. The second prediction result can be the probability value of the ship software of the target ship failing caused by the third sample data in each dimension output by the random forest regression model to be trained. The second error data can be used to represent the degree of difference between the second prediction result and the theoretical fault result corresponding to the third sample data.

[0075] Specifically, determine the second sample data that is not used during the construction of each decision tree in the random forest regression model to be trained, input the second sample data into the corresponding decision tree, and obtain the first output results of at least one decision tree. Perform a summation process on the first output results of each decision tree to obtain a summation result. Then, perform an averaging process on the summation result according to the number of decision trees that output the first output results to obtain a first prediction result. Determine first error data based on the first prediction result and the theoretical fault result corresponding to the second sample data.

[0076] Perform data augmentation on the second sample data in the second training sample to obtain third sample data corresponding to the second sample data. Determine a third training sample based on the theoretical fault result corresponding to the third sample data in the original training sample and the third sample data. Input the third sample data in the third training sample into at least one decision tree of the random forest regression model to be trained to obtain a second prediction result. Determine second error data based on the second prediction result and the theoretical fault result of the third training sample. Correct the model parameters of the random forest regression model to be trained according to the first error data and the second error data to obtain a trained random forest regression model.

[0077] Exemplarily, in combination with the above example, when constructing each decision tree of the random forest regression model to be trained using the first sample data, each decision tree corresponds to a part of the original sample data that is not used for constructing the decision tree. This part of the original sample data can be used as out-of-bag data, that is, the second sample data. Input the second sample data into the random forest regression model to be trained to determine the first error data corresponding to the second sample data. For example, if the second sample data is X j , and X j is not used for constructing the m-th decision tree and the (m + 1)-th decision tree, then X j can be used to train the m-th decision tree and the (m + 1)-th decision tree. Input the second sample data X j into the m-th decision tree and the (m + 1)-th decision tree respectively to obtain the first output result of the m-th decision tree and the first output result of the (m + 1)-th decision tree. Determine the average output result, that is, the first prediction result, according to the two first output results. Determine the first error data based on the first prediction result and the theoretical fault result corresponding to the second sample data X j . The historical ship software fault data x j corresponding to the i-th dimension in the second sample data X iPerform random shuffling or noise addition to obtain the third sample data. Input the third sample data into the m-th and (m + 1)-th decision trees of the random forest regression model to be trained to determine the corresponding second error data. According to the first error data and the second error data, the importance of the historical ship software fault data in the i-th dimension in the random forest regression model can be determined. If the difference between the first error data and the second error data is small, it indicates that the historical ship software fault data in the i-th dimension has little impact on the random forest regression model to determine the fault prediction result; correspondingly, if the difference between the first error data and the second error data is large, it indicates that the historical ship software fault data in the i-th dimension has a greater impact on the random forest regression model to determine the fault prediction result. Then, according to the degree of difference between the first error data and the second error data, the model parameters of the random forest regression model to be trained can be adjusted to obtain a random forest regression model for determining ship faults.

[0078] Optionally, during the training process of the random forest regression model to be trained, the third training sample can be determined as follows: for the historical ship software fault data in multiple dimensions in the second training sample, add random Gaussian noise to the historical ship software fault data in the current dimension to obtain the perturbed data corresponding to the current dimension; use the perturbed data corresponding to the current dimension and the historical ship software fault data in other dimensions except the current dimension as the third sample data, and determine the theoretical fault result corresponding to the third sample data; determine the third training sample according to the third sample data and the theoretical fault result corresponding to the third sample data.

[0079] Among them, random Gaussian noise can be understood as noise data that follows a Gaussian distribution. The perturbed data can be the historical ship software fault data in the current dimension after adding random Gaussian noise.

[0080] Specifically, for the historical ship software fault data in multiple dimensions in the second training sample, add random Gaussian noise to the historical ship software fault data in a certain dimension to obtain the perturbed data corresponding to that dimension. Determine the third sample data according to the perturbed data in that dimension and the historical ship software fault data in other dimensions except that dimension. Determine the third training sample according to the third sample data and the theoretical fault result corresponding to the third sample data, so as to determine the influence degree of the data in each dimension on the random forest regression model to determine the fault prediction result based on the third training sample and the second training sample.

[0081] It should be noted that for multiple dimensions, the above-mentioned processing of adding random Gaussian noise can be separately performed on the historical ship software fault data of each dimension of the second sample data to obtain multiple third sample data. In each third training sample, only the historical ship software fault data of one dimension has undergone data augmentation processing. Based on the multiple third sample data and the corresponding theoretical fault results, the second error data corresponding to each dimension is determined.

[0082] Optionally, during the training process of the random forest regression model to be trained, the method of correcting the model parameters of the random forest regression model to be trained through the first error data and the second error data can be: based on the first error data and the second error data, determine the target error data corresponding to each dimension; based on the target error data corresponding to each dimension, determine the weight coefficient corresponding to each dimension, and based on the weight coefficient, correct the model parameters of the random forest regression model to be trained to obtain the trained random forest regression model.

[0083] Among them, the first error data can be used to characterize the degree of difference between the first prediction result and the theoretical fault result corresponding to the second sample data. For example, if the second sample data contains historical ship software fault data of 3 dimensions, then the theoretical fault result corresponding to the second sample data contains the theoretical probability values of the historical ship software fault data of each of these 3 dimensions causing the target ship to malfunction. The first prediction result contains the predicted probability values output by the random forest regression model based on the second sample data for the historical ship software fault data of each of these three dimensions causing the target ship to malfunction. Correspondingly, the second error data can be used to characterize the degree of difference between the second prediction result and the theoretical fault result corresponding to the third sample data. The target error data can be used to characterize the degree of difference between the first error data and the second error data corresponding to each dimension. The weight coefficient can be a coefficient used to characterize the importance of each dimension.

[0084] Specifically, when performing data augmentation processing on the second training sample, the data augmentation processing can be separately performed on the historical ship software fault data of each dimension in the second sample data of the second training sample, so as to determine the third training sample corresponding to this dimension through the historical ship software fault data after data augmentation processing of this dimension and the historical ship software fault data of other dimensions that have not undergone data augmentation processing. Based on this, multiple third training samples can be obtained. Correspondingly, for multiple third training samples, the second error data corresponding to each dimension can be obtained. For the second error data corresponding to multiple dimensions, perform a subtraction operation on the first error data and the second error data corresponding to the current dimension to obtain a difference result, and based on this difference result and the number of all decision trees in the random forest regression model, determine the target error data corresponding to the current dimension.

[0085] Sort the target error data corresponding to each dimension, and determine the weight coefficient corresponding to each dimension according to the sorting result, so as to correct the model parameters of the random forest regression model to be trained, and obtain the trained random forest regression model. It should be noted that if the target error data corresponding to the current dimension is small, the weight coefficient corresponding to the current dimension is also small.

[0086] Exemplarily, in combination with the above example, for the second sample data X j the historical ship software fault data x i corresponding to the i-th dimension in is randomly shuffled or noise-added to obtain the third sample data. Input the third sample data into the random forest regression model to be trained to obtain the second error data. Determine the difference result according to the first error data and the second error data corresponding to the i-th dimension, and determine the target error data according to the difference result and the number of all decision trees in the random forest regression model. Optionally, the target error data can be determined by the following function.

[0087]

[0088] where n represents the number of all decision trees in the random forest regression model, a = 1 represents the first decision tree in the random forest regression model, and OOB shuffled (X j ) represents the second error data corresponding to the i-th dimension in the second sample data X j , and OOB orig represents the first error data.

[0089] Through the above processing, determine the target error data corresponding to data in multiple dimensions such as meteorological data, equipment status data, equipment environment data, geographical location data, and operation behavior data. Perform a descending sorting process on the target error data corresponding to multiple dimensions to obtain the sorting result. According to the sorting result, determine the weight coefficient corresponding to each dimension, so as to correct the model parameters of the random forest regression model to be trained based on the weight coefficient, and obtain the trained random forest regression model. Optionally, the dimensions corresponding to the target error data below the preset error threshold can also be removed according to the preset error threshold, and the above process is repeated until the final N dimensions are obtained. When applying the random forest regression model for determining ship faults subsequently, only the first data corresponding to the corresponding N dimensions needs to be obtained, so as to determine which of the first data in specific dimensions leads to the target ship software fault based on the first data in the N dimensions, and thus perform automatic operation and maintenance processing on the ship software of the target ship according to the result.

[0090] In the technical solution of this embodiment, multiple original training samples are obtained, the original sample data of the multiple original training samples is sampled according to a preset sampling algorithm to obtain first sample data, and based on the first sample data and the corresponding theoretical fault results, first training samples are obtained. Based on the first training samples, at least one decision tree is determined using a preset decision tree algorithm, so as to obtain a to-be-trained random forest regression model based on the at least one decision tree. For at least one decision tree of the to-be-trained random forest regression model, the original sample data that has not been used to construct the decision tree is used as second sample data, and based on the second sample data and the corresponding theoretical fault results, second training samples are obtained. The to-be-trained random forest regression model is trained based on the second training samples to obtain a trained random forest regression model, and the trained random forest regression model is determined as the random forest regression model for determining ship faults. Based on this, the construction and training of the random forest regression model are realized, which facilitates the subsequent deployment of the trained random forest regression model for determining ship faults in the target ship, so as to timely determine the factors causing ship faults when a fault in the ship software of the target ship is detected, improving the operation and maintenance efficiency of the ship software. And it further ensures the stability and security of the operation of the ship software of the target ship in the case where there is no satellite network communication and the operation and maintenance personnel cannot remotely perform operation and maintenance on the ship software.

[0091] Embodiment III

[0092] Figure 4 It is a schematic structural diagram of a ship software operation and maintenance device provided in Embodiment III of the present invention. As Figure 4 shown, the device includes: a data acquisition module 310, a fault determination module 320, and a ship software maintenance module 330.

[0093] The data acquisition module 310 is configured to acquire first data in multiple dimensions within a preset time period before the ship software fault moment of the target ship. Among them, the first data in multiple dimensions includes at least two of the meteorological data corresponding to the target ship, the equipment status data of the target ship, the environment data where the equipment of the target ship is located, the geographical location data of the target ship, and the operation behavior data of the target ship; the fault determination module 320 is configured to process the first data based on a pre-trained random forest regression model for determining ship faults to obtain a fault prediction result, where the fault prediction result includes the fault evaluation attributes corresponding to the first data in each dimension, and the random forest regression model is a model deployed in the target ship; the ship software maintenance module 330 is configured to generate ship software operation and maintenance information based on the fault prediction result and the first data, so as to perform software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information.

[0094] The technical solution of this embodiment is to obtain the first data in multiple dimensions within a preset time period before the moment of ship software failure of the target ship, and process the first data through a pre-trained random forest regression model for determining ship faults deployed on the target ship to determine the fault evaluation attributes corresponding to the first data in each dimension, that is, obtain the fault prediction result, and determine the probability that the first data in each dimension causes the ship software of the target ship to fail. According to the fault prediction result, the factors affecting the normal operation of the ship software can be determined, solving the problems in the prior art that it is difficult to determine the specific fault cause and the manual operation and maintenance cost is too high caused by manual remote operation and maintenance of the ship software. Through the fault prediction result and the first data, ship software operation and maintenance information is generated to perform software maintenance on the ship software of the target ship based on the ship software operation and maintenance information. The present invention solves the problems such as high operation and maintenance cost, low ship software operation and maintenance efficiency, and difficult fault cause location caused by manual remote operation and maintenance of ship software. By analyzing the first data in multiple dimensions to determine the fault prediction result, the accuracy of fault cause location is ensured, and more comprehensive software maintenance processing can be provided for the ship software of the target ship. By using the random forest regression model deployed on the target ship for fault detection, it is realized that the ship software anomaly can be automatically processed without manual remote operation and maintenance of the ship software, improving the operation and maintenance efficiency of the ship software, and ensuring the stability and security of the operation of the ship software of the target ship in the case where there is no satellite network communication and the operation and maintenance personnel cannot remotely operate and maintain the ship software.

[0095] Based on the above embodiments, optionally, the device further includes: a random forest regression model training module, which includes: an original training sample acquisition unit for acquiring a plurality of original training samples, where the original training samples include original sample data and theoretical fault results corresponding to the original sample data, and the original sample data includes historical ship software fault data in multiple dimensions; a first training sample determination unit for sampling the original sample data of the plurality of original training samples according to a preset sampling algorithm to obtain first sample data, and obtaining a first training sample according to the first sample data and the theoretical fault results corresponding to the first sample data; a random forest regression model construction unit for determining at least one decision tree based on the first training sample by using a preset decision tree algorithm, so as to obtain a to-be-trained random forest regression model based on the at least one decision tree; a second training sample determination unit for, for at least one decision tree of the to-be-trained random forest regression model, taking the original sample data not used for constructing the decision tree as second sample data, and obtaining a second training sample based on the second sample data and the theoretical fault results corresponding to the second sample data, where the second sample data includes historical ship software fault data in multiple dimensions; a model training unit for training the to-be-trained random forest regression model based on the second training sample to obtain a trained random forest regression model, and determining the trained random forest regression model as the random forest regression model for determining ship faults.

[0096] Optionally, the model training unit includes: a first output result determination subunit for inputting the second sample data of the second training sample into at least one decision tree of the to-be-trained random forest regression model to obtain a first output result of each decision tree, where the first output result includes a fault evaluation attribute corresponding to the historical ship software fault data in each dimension; a first error data determination subunit for determining a first prediction result based on the first output result of each decision tree and the number of decision trees, and determining first error data based on the first prediction result and the theoretical fault results corresponding to the second sample data; a third training sample determination subunit for performing data augmentation processing on the second training sample to obtain a third training sample, where the third training sample includes third sample data corresponding to the second sample data and theoretical fault results corresponding to the third sample data; a second error data determination subunit for inputting the third sample data into at least one decision tree of the to-be-trained random forest regression model to obtain a second prediction result, and obtaining second error data according to the second prediction result and the theoretical fault results corresponding to the third sample data; a model parameter correction subunit for correcting the model parameters of the to-be-trained random forest regression model based on the first error data and the second error data to obtain a trained random forest regression model.

[0097] Optionally, a third training sample determination subunit is configured to add random Gaussian noise to the historical ship software fault data in the current dimension for the historical ship software fault data in multiple dimensions in the second training sample, so as to obtain the perturbed data corresponding to the current dimension; use the perturbed data corresponding to the current dimension and the historical ship software fault data in other dimensions except the current dimension as the third sample data, and determine the theoretical fault result corresponding to the third sample data; determine the third training sample according to the third sample data and the theoretical fault result corresponding to the third sample data.

[0098] Optionally, a model parameter correction subunit is configured to determine the target error data corresponding to each dimension based on the first error data and the second error data; determine the weight coefficient corresponding to each dimension based on the target error data corresponding to each dimension, and correct the model parameters of the random forest regression model to be trained based on the weight coefficient, so as to obtain the trained random forest regression model.

[0099] Optionally, a data acquisition module is configured to, when detecting that a target ship has a fault, acquire the original data in multiple dimensions within a preset time period before the ship software fault moment of the target ship; perform data preprocessing on the original data to obtain the first data associated with the ship software fault of the target ship, where the data preprocessing includes at least one of data cleaning processing, outlier processing, and data normalization processing based on preset data screening conditions.

[0100] Optionally, the apparatus further includes: a model parameter update module, configured to perform an update process on the model parameters of the random forest regression model for determining ship faults based on the operation and maintenance logs corresponding to the ship software operation and maintenance information, so as to obtain the updated random forest regression model, and process the next first data based on the updated random forest regression model to obtain the corresponding fault prediction result.

[0101] The ship software operation and maintenance apparatus provided by the embodiments of the present invention can execute the ship software operation and maintenance method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.

[0102] Embodiment 4

[0103] Figure 5FIG. 0 is a schematic structural diagram of an electronic device provided in Embodiment 4 of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as, for example, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as, for example, personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely exemplary and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0104] As Figure 5 shown, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. The memory stores a computer program executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0105] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0106] The processor 11 may be various general-purpose and / or special-purpose processing components having processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the ship software operation and maintenance method.

[0107] In some embodiments, the ship software operation and maintenance method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the ship software operation and maintenance method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the ship software operation and maintenance method by any other suitable means (e.g., by means of firmware).

[0108] Various embodiments of the systems and techniques described above in this document may be implemented in digital electronic circuitry, integrated circuit systems, field-programmable gate arrays (FPGA), application specific integrated circuits (ASIC), application specific standard products (ASSP), systems on a chip (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that may be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a special-purpose or general-purpose programmable processor that may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0109] The computer program for implementing the ship software operation and maintenance method of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer programs are executed by the processor, the functions / operations specified in the flowchart and / or block diagram are implemented. The computer programs may be executed entirely on the machine, partially on the machine, as a stand-alone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0110] Embodiment 5

[0111] Embodiment 5 of the present invention further provides a computer-readable storage medium storing computer instructions for causing a processor to execute a ship software operation and maintenance method, the method including:

[0112] Obtain first data of a target ship in multiple dimensions within a preset duration before the moment of ship software failure. Among them, the first data in multiple dimensions includes at least two of the following: meteorological data corresponding to the target ship, equipment status data of the target ship, environmental data where the equipment of the target ship is located, geographical location data of the target ship, and operation behavior data of the target ship. Process the first data based on a pre-trained random forest regression model for determining ship faults to obtain a fault prediction result. The fault prediction result includes a fault evaluation attribute corresponding to the first data in each dimension. The random forest regression model is a model deployed in the target ship. Generate ship software operation and maintenance information based on the fault prediction result and the first data, so as to perform software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information.

[0113] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0114] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, speech input, or tactile input).

[0115] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend, middleware, or frontend components. The components of the system can be interconnected to each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.

[0116] The computing system can include clients and servers. The clients and servers are generally remote from each other and typically interact through a communication network. The client-server relationship is created by computer programs running on respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, solving the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0117] It should be understood that various forms of the processes shown above can be used, steps can be reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is made herein.

[0118] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for ship software operation and maintenance, characterized in that, Including: Obtain first data in multiple dimensions within a preset time period before the moment of ship software failure for the target ship. Among them, the first data in the multiple dimensions includes at least two of the following: meteorological data corresponding to the target ship, equipment status data of the target ship, environmental data where the equipment of the target ship is located, geographical location data of the target ship, and operation behavior data of the target ship; Process the first data based on a pre-trained random forest regression model for determining ship faults to obtain a fault prediction result. Among them, the fault prediction result includes a fault evaluation attribute corresponding to the first data in each dimension, and the random forest regression model is a model deployed in the target ship; Generate ship software operation and maintenance information based on the fault prediction result and the first data, so as to perform software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information.

2. The method according to claim 1, wherein The method further includes: Obtain a plurality of original training samples. Among them, the original training samples include original sample data and a theoretical fault result corresponding to the original sample data, and the original sample data includes historical ship software fault data in the multiple dimensions; Perform sampling processing on the original sample data of the plurality of original training samples according to a preset sampling algorithm to obtain first sample data, and obtain a first training sample according to the first sample data and the theoretical fault result corresponding to the first sample data; Based on the first training sample, use a preset decision tree algorithm to determine at least one decision tree, so as to obtain a random forest regression model to be trained based on the at least one decision tree; For at least one decision tree of the random forest regression model to be trained, use the original sample data that has not been used to construct the decision tree as second sample data, and obtain a second training sample based on the second sample data and the theoretical fault result corresponding to the second sample data. Among them, the second sample data includes historical ship software fault data in multiple dimensions; Perform training processing on the random forest regression model to be trained based on the second training sample to obtain a trained random forest regression model, and determine the trained random forest regression model as the random forest regression model for determining ship faults.

3. The method according to claim 2, wherein The performing training processing on the random forest regression model to be trained based on the second training sample to obtain a trained random forest regression model includes: Input the second sample data of the second training sample into at least one decision tree of the random forest regression model to be trained to obtain a first output result of each decision tree. Among them, the first output result includes a fault evaluation attribute corresponding to historical ship software fault data in each dimension; Determine a first prediction result based on the first output result of each decision tree and the number of decision trees, and determine first error data based on the first prediction result and the theoretical fault result corresponding to the second sample data; Perform data augmentation on the second training sample to obtain a third training sample, where the third training sample includes third sample data corresponding to the second sample data and theoretical fault results corresponding to the third sample data; Input the third sample data into at least one decision tree of the random forest regression model to be trained to obtain a second prediction result, and obtain second error data based on the second prediction result and the theoretical fault results corresponding to the third sample data; Based on the first error data and the second error data, correct the model parameters of the random forest regression model to be trained to obtain a trained random forest regression model.

4. The method according to claim 3, wherein The performing data augmentation on the second training sample to obtain a third training sample includes: For the historical ship software fault data in multiple dimensions in the second training sample, add random Gaussian noise to the historical ship software fault data in the current dimension to obtain the perturbed data corresponding to the current dimension; Use the perturbed data corresponding to the current dimension and the historical ship software fault data in other dimensions except the current dimension as the third sample data, and determine the theoretical fault results corresponding to the third sample data; Determine the third training sample according to the third sample data and the theoretical fault results corresponding to the third sample data.

5. The method according to claim 3, wherein The based on the first error data and the second error data, correcting the model parameters of the random forest regression model to be trained to obtain a trained random forest regression model includes: Based on the first error data and the second error data, determine the target error data corresponding to each dimension; Based on the target error data corresponding to each dimension, determine the weight coefficient corresponding to each dimension, and based on the weight coefficient, correct the model parameters of the random forest regression model to be trained to obtain a trained random forest regression model.

6. The method according to claim 1, characterized in that, The obtaining first data in multiple dimensions within a preset time period before the ship software fault moment of the target ship includes: When it is detected that the target ship has a fault, obtain the original data in multiple dimensions within a preset time period before the ship software fault moment of the target ship; Perform data preprocessing on the original data to obtain first data associated with the ship software fault of the target ship, where the data preprocessing includes at least one of data cleaning processing, outlier processing, and data normalization processing based on preset data screening conditions.

7. The method according to claim 1, wherein After performing software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information, the method further includes: Based on the operation and maintenance log corresponding to the ship software operation and maintenance information, perform update processing on the model parameters of the random forest regression model for determining ship faults to obtain an updated random forest regression model, so as to process the next first data based on the updated random forest regression model to obtain the corresponding fault prediction result.

8. A ship software operation and maintenance device, characterized in that Includes: A data acquisition module, configured to acquire first data in multiple dimensions within a preset time period before the moment of ship software failure of a target ship, wherein the first data in the multiple dimensions includes at least two of the meteorological data corresponding to the target ship, the equipment status data of the target ship, the environmental data where the equipment of the target ship is located, the geographical location data of the target ship, and the operation behavior data of the target ship; A fault determination module, configured to process the first data based on a pre-trained random forest regression model for determining ship faults to obtain a fault prediction result, wherein the fault prediction result includes a fault evaluation attribute corresponding to the first data in each dimension, and the random forest regression model is a model deployed in the target ship; A ship software maintenance module, configured to generate ship software operation and maintenance information based on the fault prediction result and the first data, so as to perform software maintenance processing on the ship software of the target ship based on the ship software operation and maintenance information.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor, so that the at least one processor can execute the ship software operation and maintenance method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, and the computer instructions are used to implement the ship software operation and maintenance method according to any one of claims 1-7 when executed by a processor.

Images